Causal Uncertainty in Chinese Medical Liability

Transcription

1 Causal Uncertainty in Chinese Medical Liability Xiaowei Yu 1 Defining Causation in Chinese Tort Law Causation is an essential criterion of liability in Chinese tort law. In China, the statue directly applicable to tort claims was the General Principles of the Civil Law (1986) 1 (hereinafter GPCL) until it was replaced by the currently applicable Tort Liability Law (2009) 2 (hereinafter TLL) on July 1, Similar to many European jurisdictions (Spier and Haazen, 2000; von Bar, 2000, p. 435), China recognizes causation as a requirement of tort liability, as evidenced by reading the general clauses of tort liability - art. 6 para. 1 of the TLL. It reads: One who is at fault for infringement upon a civil right or interest of another person shall be subject to the tort liability, which is roughly equivalent to its predecessor - art. 106 para. 2 of the GPCL. 3 The word infringement ( 侵害 ) unequivocally indicates that there must be a causal link between fault and damage in order for a tortfeasor to be held liable. Causation is also a requirement of medical malpractice liability. Medical malpractice is treated as a special type of tort by the TLL, which is regulated by Chapter 7 under the title of Liability for Iatrogenic Injury ( 医疗损害责任 ). Its specialness lies in the fact that the locus of liability is placed solely on institutional health care providers (art. 54 TLL), the fact that in some cases 4 a statutory rule of presumption of fault may apply (art. 58 TLL), and the fact that strict liability applies to cases where iatrogenic injuries were caused by the defect of any drug, medical disinfectant or medical instrument (art. 59 TLL). More remarkably, art. 54 of the TLL functions as an independent general clause for medical malpractice liability in the sense that it is sufficient to base malpractice claims on art. 54 without citing art. 6 para. 1. Modelled on art. 6 para. 1, art. 54 provides in more detail: Where a patient sustains any harm during diagnosis and treatment, if the medical institution or any of its medical staff is at fault, the medical institution shall assume the compensatory liability. The word sustain ( 受到 ) also implicates a requirement of causation between fault on the part of health care providers and harm suffered by patients. Despite the fact that causation plays an essential role in tort law, none of the provisions in the TLL or the GPCL has ever defined this term. As a matter of fact, causation theories are largely developed by legal scholars and judges in China. During the last two decades of the 20 th century in China, the dominant idea about the causation rule in tort law is the so-called Xiaowei Yu: PhD candidate at Maastricht University (Xiaowei.yu@maastrichtuniversity.nl). 1 General Principles of the Civil Law 民法通则 (promulgated by the Nat l People s Cong., Apr. 12, 1986, effective Jan. 1, 1987, amended Aug. 27, 2009) 1986(2) SUP. PEOPLE S CT. GAZ. 15, translated in (Lawinfochina). 2 Tort Liability Law 侵权责任法 (promulgated by the Standing Comm. Nat l People s Cong., Dec. 26, 2009, effective Jul. 1, 2010) 2010(1) STANDING COMM. NAT L PEOPLE S CONG. GAZ. 4, translated in (Lawinfochina). 3 Art. 106 para. 2 of the GPCL provides, Citizens and legal persons who through their fault encroach upon state or collective property, or the property or person of other people shall bear civil liability. 4 Art. 58 of the TLL provides that a medical institution shall be presumed to be at fault for any iatrogenic injury if the institution or its medical staff members are proven to have (1) violated any statutorily required procedure or standard for diagnosis and treatment, (2) concealed or refused to provide relevant medical record, or (3) forged, tampered or destroyed any relevant medical record. 1

2 necessity 5 theory ( 必然因果关系说 ). According to this theory, causation only exists where there is an inner, intrinsic and inevitable link ( 内在的 本质的 必然的联系 ) between the injurious act and damage (Zhao, 2008), i.e. where the movement from the injurious act to damage indicates a trend that is inevitable and absolutely certain according to objective natural law (Jiang, 2010, p. 223). This approach was claimed to be based on dialectical materialism and influenced by the then Soviet law (Zhao, 2008), which is also similar to Kant s view that D necessarily follows from A or Hume s view that in every instance D always follows A, if A is said to be the cause of D (Young, Faure and Fenn, 2004). Put another way, the necessity theory requires that A is a sufficient condition for D to occur and the causal trend is certain. It does not matter whether A is a necessary condition for D or not. However, in modern tort cases, especially those involving industrial and medical accidents, it is extremely difficult for victims to prove that D necessarily follows from A. For instance, many treatment outcomes are random rather than predetermined and statistical regularity such as the cure rate is based on probabilities (Zhao, 2008). Therefore, the necessity theory considerably limits both the possibility of establishing tort liability and the scope of compensation, being extremely disadvantageous to the protection of the patient s interests. The stringent requirement of the necessity theory has been challenged and relaxed. Since the 1990 s, the necessity theory had been widely criticized by the academia. The first scholar challenged the orthodox necessity theory was perhaps Liang (1989). In his commentary on Zhang & Zhang v. Zhang, 6 which was published on the Supreme People s Gazette in 1989, Liang criticized the necessity theory on the grounds that to understand the inner, intrinsic and inevitable link between objective things is beyond the capacity of most judges (ibid.). What judges can do is to decide cases mainly in accordance with the common norms of social life, the concepts of equity and justice, boni mores and normal human feelings (ibid.). After the criticism, Liang advocated the use of the adequacy theory, which was also implicitly employed in Zhang & Zhang v. Zhang. In the case, the victim Mr. Zhang had his left inner ankle wounded due to an accident on a construction site. The hospital appropriately treated him but failed to save his life. Mr. Zhang later died of septicopyemia, septic shock and multiple organ failure, which was not due to any fault on the part of the hospital. Had the court adhered to the necessity theory, the defendant-employer would not have been held liable because death does not necessarily follow a leg wound. On the contrary, the court nonetheless confirmed that the victim s death could only have been attributed to the industrial injury. Although the court had not expressly mentioned based on which theory the causation was established, Liang (1989) argued that the decision could only be explained on the basis of the adequacy theory ( 相当因果关系说 ) rather than the necessity theory. According to Liang, causation can be established if, based on currently available knowledge and experience, an ordinary person agrees that there may be the same harm under the same circumstance. Liang s commentary on Zhang & Zhang v. Zhang marked a turning point in Chinese causation theories in tort law. Although Zhang & Zhang v. Zhang was not adjudicated by the 5 The word necessity here only refers to a situation that is inevitable, which should not be confused with the legal defense of necessity. 6 Lianqi Zhang & Guoli Zhang v. Xuezhen Zhang 张连起 张国莉诉张学珍损害赔偿纠纷案, 1989(1) SUP. PEOPLE S CT. GAZ. 23 (Tianjin Tanggu Dist. People s Ct. 1988) (China). 2

3 Supreme People s Court, 7 the publication of the case on its official gazette clearly indicated the Supreme Court s preference for the adequacy theory. Later, more and more scholars proposed to introduce the adequacy theory from German law into Chinese tort law (Zhu, 2004; Yang, 2010, p. 98; Cheng, 2011, p. 181; Wang, 2011, p. 375). Currently, it seems that the adequacy theory tends to play a more important role in tort decisions (Cheng, 2011, p. 182). In spite of vital importance, neither the TLL/GPCL nor any judicial interpretation 8 has ever provided for the definition of causation. Nevertheless, some clues to its definition can be discovered at the local level. For instance, two provincial high courts expressly define the adequacy theory by issuing judicial guides 9 within their own jurisdictions. In 2005, the Shanghai High People s Court enacted the Guide to the Trial of Medical Malpractice Cases (2005) 10 (hereinafter Shanghai Guide). Art. 13 of the Shanghai Guide provides that the determination of causation shall be based on the adequacy theory an adequate causation exists between a negligent treatment and an injury if the treatment is the conditio sine qua non of the injury and if the treatment substantially increases the objective probability of the occurrence of the injury. Likewise, the Jiangsu High People s Court issued the Guide to the Trial of Tort Compensation Cases (2011) 11 (hereinafter Jiangsu Guide), art. 5-1 of which provides for the definition of the adequacy theory as follows: There is causation between an action and an injury, if the injury will not occur but for the action, and it is reasonably likely that the presence of the action will result in the injury; there is no causation, if the injury will not occur but for the action, but the injury will normally not occur even in the presence of the action. Combining the foregoing two guides, it is evident that the adequacy theory bases causation on two essential tests: (1) the but for or conditio sine qua non test, and (2) the adequacy an increased objective probability or reasonable likelihood. This two-step approach is extremely similar to that of German tort law (Magnus, 2000, pp ). Currently, however, there is no evidence that a third step policy considerations such as those associated with the protective purpose of the violated rule has been or will be adopted by Chinese courts yet. In 7 The Chinese court system in the mainland consists of local courts, special courts and the Supreme People s Court, with all the first two subject to the supervision of the latter. There are 32 provincial jurisdictions, each of which has a three-level court system: district people s courts at the grassroots-level, intermediate people s courts at the municipal level and high people s courts at the provincial level. Normally, most civil cases of the first instance are tried before district courts and those of the second instance before intermediate courts. The second instance is the final instance. Therefore, the vast majority of civil cases will not reach the Supreme Court. Nevertheless, the Supreme Court has the power to overrule or retry any erroneous local adjudications if the Supreme Court deems it necessary. The Supreme Court also regularly select exemplary local decisions and publish them in order for all local courts nationwide to follow. For more detailed information about China s court system, see China s Judiciary, available at < > (accessed November 19, 2015). For the recently established Chinese Guiding Case System, see Deng (2015). 8 In China, the Supreme People s Court is empowered by Constitution to enact judicial interpretations concerning the application of law. Judicial interpretations can be directly cited in decisions and are as applicable and enforceable as national laws promulgated by the People s Congress or its Standing Committee. 9 Local judicial guides issued by high or intermediate courts are not judicial interpretations. They are by no means applicable rules and cannot be directly cited in decisions. In practice, nevertheless, these local guides are highly influential in local judges application of law in the sense that cases of the first instance may be overruled by appellate courts if trial judges do not strictly follow relevant guides issued by the appellate courts. In other words, local judicial guides represent opinions of appellate courts. 10 Guide to the Trial of Medical Malpractice Cases 医疗过失赔偿纠纷案件办案指南 (promulgated by the Shanghai High People s Ct., Sep. 20, 2005, effective Sep. 20, 2005) (Chinalawinfo), translated by the author. 11 Trial Guide to the Trial of Tort Compensation Cases 侵权损害赔偿案件审理指南 (promulgated by the 1st Civil Trial Chamber Jiangsu High People s Ct., 2011, effective 2011) (Chinalawinfo), translated by the author. 3

4 practice, many Chinese courts also refer to the first step as causation in fact ( 事实上的因果关系 ) and the second one as causation in law ( 法律上的因果关系 ). It should be noted, however, that the Chinese version of causation in law is different from its counterpart remoteness or legal cause or proximate cause in the common law (Rogers, 2000, p. 40). Whereas the test for the former is the probability-based adequacy, the leading test for the latter is foreseeability (ibid.). 2 Burden and Standard of Proof 2.1 The Burden of Proof In China, the burden of proof in civil cases rests on the plaintiff as a matter of principle unless otherwise provided by the law. According to art. 2 para. 2 of the Provisions on Evidence in Civil Procedures (2001) 12 (hereinafter PECP), when the party that bears the burden of proof fails to produce sufficient evidence to support his allegations, he will face unfavorable consequences. In medical malpractice cases, the issue of the burden of proof is complex and controversial. For the period , the GPCL, in conjunction with the Measures for the Handling of Medical Accidents (1987) 13 (hereinafter MHMA), were applicable to the resolution of medical malpractice claims. There was no special provision on the burden of proof regarding malpractice claims and it was the plaintiff-patient that bore the burden of proving all elements of malpractice liability during that period. This general allocation of the burden of proof was dramatically reversed for medical malpractice liability in the following decade ( ). As of 2001, according to art. 4 para. 1 sub-para. 8 of the PECP, the defendant-hospital began to bear the burden of proving the absence of both fault and causation while the plaintiffpatient only had to prove that he had suffered iatrogenic injuries in the defendant-hospital. This nearly full reversal of the burden of proof was criticized especially for two reasons (Shao, 2013). First, it might encourage frivolous actions. Patients will be more willing to file malpractice claims if they bear little burden of proof. Second, it might induce providers to practice defensive medicine. Providers might take excessive precautions in respond to increased risk of losing a malpractice lawsuit due to failure to disprove causation and fault. As a result, in 2010, the reversal rule was reversed once more by art. 54 of the TLL back to the original state. For the time being, it is normally the plaintiff-patient that has to prove the presence of harm, negligence and causation. There is still one exception to the general allocation of the burden of proof the defendant-hospital is presumed to be at fault under one of the three circumstances 14 provided for at art. 58 of the TLL. It appears that art. 58 of the 12 Some Provisions on Evidence in Civil Procedures 关于民事诉讼证据的若干规定 (promulgated by the Sup. People s Ct., Dec. 21, 2001, effective Apr. 1, 2002) 2002(1) SUP. PEOPLE S CT. GAZ. 22, translated in (Lawinfochina). 13 Measures for the Handling of Medical Accidents 医疗事故处理办法 (promulgated by the St. Council, Jun. 29, 1987, effective Jun. 29, 1987, annulled Sep. 1, 2002) 1987(18) ST. COUNCIL GAZ. 614, translated by the author. 14 See supra. note 4. 4

5 TLL only conditionally reverses the burden of proving fault rather than causation. However, when reading the last two sub-paragraphs of art. 58 closely, it is questionable whether causation is presumed as well. When the plaintiff-patient is denied access to his medical records or the records are forged, tampered or destroyed, it is virtually impossible for the plaintiff-patient to prove the existence of causation, for medical records are the key evidence in medical malpractice cases (Ye, 2012). There is some evidence showing that courts tend to hold the defendant-hospital liable in cases where there are some obstacles to obtaining authentic medical records. For example, in Zhang v. Suiping Hospital, 15 the victimized patient s medical records were forged. The Suiping Court presumed the defendant-hospital to be at fault in accordance with art. 58 sub-para. 3 of the TLL, and held the defendant directly liable for the patient s death according to some legal provisions. The court did not clarify what some legal provisions were. Given the fact that the court only cited art. 58 sub-para. 3 in the decision, it is reasonable to assume that the court applied sub-para. 3 by analogy to the determination of causation. 2.2 The Standard of Proof Common law countries and civil law countries approach the standard of proof differently. In common law systems, whereas a very high standard beyond a reasonable doubt applies to criminal proceedings, a relatively low more-likely-than-not (at least 51%) standard, termed preponderance of evidence in the US and balance of probabilities in the UK, is used in civil cases (Clermont and Sherwin, 2002; Taruffo, 2003; Wright, 2011; Schweizer, 2013). Unlike common law, it is commonly argued that there is no distinction between standards of proof applying to civil and criminal proceedings in civil law systems (ibid.). In France, the subjective persuasion of a single judge intime conviction is required (ibid.). In Germany, the standard is interpreted as a full judicial conviction in the form of a degree of certainty that silences doubts for practical purposes, even if it does not eliminate them entirely (Stauch, 2008). The standard of proof in Chinese law, however, seems as uncertain as the causal link between wrongful treatment and harm done. Generally speaking, the civil procedure and proof system in China follows the continental tradition and is basically inquisitorial. The Civil Procedure Law (2012) 16 (hereinafter CPL), however, never specifies what the standard is. According to scholarly opinions, the same standard of clear facts, unquestionable and sufficient evidence ( 事实清楚, 证据确实充分 ) applied to both criminal and civil proceedings in the 1990s (Wang, 1999). The application of this very high standard to civil cases has been subject to fierce criticism and a general consensus among scholars has been reached that a less strict standard of proof should be applied to civil cases (Zhang, 2003). Consequently, the Supreme People s Court set the standard of proof for civil proceedings at art. 73 para. 1 of the PECP, which reads as follows: 15 Jiayou Zhang et al. v. Suiping County People s Hospital 张加友等诉遂平县人民医院医疗损害责任纠纷案, (Chinalawinfo) (Henan Province Suiping County People s Ct. 2013) (China). 16 Civil Procedure Law 民事诉讼法 (promulgated by the Standing Comm. Nat l People s Cong., Apr. 9, 1991, effective Arp. 9, 1991, amended Oct. 28, 2007, Aug. 31, 2012) 2012(Z2) NEW L. & REG. 3, translated in (Lawinfochina). 5

6 Where both parties concerned produces contradicting evidences to prove a same fact but neither has enough evidence to rebut the evidence of the other party, the People s court shall determine which evidence are obviously more forceful than the other evidence by taking the case into consideration, and shall affirm the evidence that are more forceful [italics added]. Although drafters of the PECP interpreted the obviously-more-forceful standard as high probability ( 高度盖然性 ) standard, they did not clarify how high the probability should be (Li, 2002). Some scholars advocate that this probability should be nine times out of ten (Li, 1999) or at least 70% (Wu, 2013). Therefore, similar to common law, Chinese law adopts a probability-oriented rather than intime conviction approach to the standard of proof in civil cases. However, the high probability standard demands a higher probability than just being more likely than not. Be that as it may, due to lack of clear definition of the high probability, Chinese judges excise their discretion to interpret the required degree of proof. A recent case study shows that many judges equate the high probability standard with the preponderance of evidence and some judges simply apply the preponderance of evidence directly even without reference to the high probability standard (ibid.). Because there are dividing interpretations of the high probability standard, the application of the standard of proof in civil cases is not unified and consistent in China. 3 Legal Approaches to Causal Uncertainty 3.1 The Multiplicity of Causal Factors In many tort cases, especially those involving medical malpractice, the adequacy theory may run short of ways in dealing with complex situation. Causation issues can be complicated in that there is usually more than one causal chain or cause that results in the same damage. For instance, injuries suffered by patients during treatment may be caused by multiple causes. Besides medical errors and system failures, injuries may result from (1) the patient s underlying medical condition or his vulnerabilities and predispositions, or (2) risks inherent in treatment beyond physicians control such as unavoidable complications (Stauch, 2008, p. 76; Yang, 2008; Wang, 2014). The but for test often fails to screen out the cause of an injury, because the injury might nevertheless occur in the absence of the faulty treatment. The adequacy test is sometimes doubtful as well due to the fact that it lacks a clear threshold. It is essential to know how courts resolve issues about causal uncertainty. 3.2 The Categorization of Causal Uncertainty Five types of typical cases may be particularly conducive to the categorization of causal 6

7 uncertainty in medical malpractice cases 17 : Case 1: P underwent heart surgery in Hospital D1. The surgery was performed by Dr. D2, who is an official medical staff member of Hospital D3. P privately employed Dr. D2 to perform the surgery without the permission of Hospital D3. Hospital D1 agreed to provide an operating room and an anesthetist and several nurses. Few days after the surgery, a piece of sterile gauze was detected inside P s chest. It is uncertain whether Dr. D2 or any medical staff member of Hospital D1 negligently dropped the piece of gauze. 18 Case 2: After being accidently wounded, P went to Hospital D1 and Hospital D2 successively. During hospitalization, both hospitals provided P with blood transfusion. After treatment, P was infected with hepatitis C. It is already proven that both D1 and D2 provided substandard blood that had been infected with the virus of hepatitis C. 19 Case 3: P is born with severe brain damage. It is uncertainty whether the damage was an inevitable outcome of her premature birth or caused due to the negligent treatment by Dr. D, the obstetrician. It is established that the probability that D s mistreatment caused the damage is 20% (or 90%). Case 4: Dr. D negligently fails to diagnose P s cancer. As P s symptoms worsen, she consults another doctor who diagnoses her cancer, but because of its advanced stage, treatment is unsuccessful, and she dies. If Dr. D had diagnosed the cancer when initially consulted, P would have had a 20% (or 90%) chance of successful life-saving treatment, which she lost. Case 5: Medical failure of Dr. D to diagnose a cancer in his patient P results in an increased size of the tumor and therefore in greater harm, but the extent of this increase in size (and consequent magnitude of harm) due to D s fault cannot be identified and established by P. Medical malpractice cases involving causal uncertainty can be grouped into two categories: those with uncertainty causes of past harm and those with indeterminate parts of harm (Gilead, Green and Koch, 2013, pp ). Cases A, B, C and D belong to the first category, and Case E to the latter. Whereas in cases of the former category it is uncertain whether D s tortious activity was the factual cause of any part of P s harm at all, in cases of the second category it is established by the required standard of proof that D tortuously caused at least some harm to P, and the causal uncertainty is which part of P s harm was caused by D (ibid., p. 15) and which part of it was caused by P s own underlying medical condition. 17 For more detailed discussion on the categorization of causal uncertainty where proportional liability may apply, see generally Gilead, Green and Koch (2013). Except for Cases 1 and 2, the remaining three cases 3, 4 and 5 are directly quoted from Gilead, Green and Koch (2013, pp. 14, 16). 18 This is a hypothetical case based on Liange Zhang et al. v. Xiuzhong Luo & Luoyang Third People s Hospital 原告张连各等诉被告罗秀忠 洛阳市第三人民医院医疗损害责任纠纷案, (Chinalawinfo) (Henan Province Song County People s Ct. 2012) (China). In China, all doctors are employed by medical institutions such as hospitals and clinics. Individual or freelance medical practice outside medical institutions is illegal. Upon approval of the medical institution where he works, however, a doctor may go out for consultation ( 外出会诊 ) to perform diagnosis and treatment within his practicing scope at another medical institution. In case an iatrogenic injury is caused by negligence on the part of the visiting doctor, it is the inviting medical institution that will be the sole defendant. See art. 2 and art. 14 of the Interim Provisions for the Administration of Doctors' Going Out for Consultation 医师外出会诊管理暂行规定 (promulgated by the Ministry of Health, Dec. 16, 2004, effective Jul. 1, 2005) 2006(8) ST. COUNCIL GAZ. 26, translated in (Lawinfochina). But if the doctor provide service at the inviting hospital without the permission of his employer-hospital, his practice is illegal and he may be held liable for his negligence (art. 6 para. 1 TLL), together with the inviting hospital (art. 54 TLL). 19 Yongqiang Hu v. Jingjiang County People s Hospital & Chinese People s Liberation Army No. 101 Hospital 胡永强与靖江市人民医院 中国人民解放军第一 o 一医院医疗损害责任纠纷案, (Chinalawinfo) (Jiangsu Province Jingjiang County People s Ct. 2014) (China). 7

8 3.3 Solutions Based on Currently Applicable Rules A. Indeterminate Tortfeasors and Multiple Sufficient Causes Case 1 involves alternative causation where which potential tortfeasor actually caused the injury is indeterminate. It is certain that P s injury was caused by the joint activity conducted by D 1 or D 2 together, but only one rather than all of them dropped the piece of gauze. This situation is similar to the well-known hunters case where it is unclear which of the negligently shooting hunters actually injured P (Gilead, Green and Koch, 2013, p. 12). The difficulty lies more in producing sufficient evidence than in causation itself (Wang, 2009, p. 189). The new Chinese TLL provides for a clear answer to this issue. Art. 10 of the TLL reads as follows: Where two or more persons engage in a conduct that endangers the personal or property safety of another person, if only the conduct of one or several of them causes harm to another person and the specific tortfeasor can be determined, the tortfeasor shall be liable; or if the specific tortfeasor cannot be determined, all of them shall be liable jointly and severally. The second clause of art. 10 TLL establishes a liability in the case of uncertain causation involving alternative defendants in China (Oliphant, 2012). Oliphant continued to argue that this approach appears to be modelled on 830(1) BGB 20 and looks similar to its counterpart in many other civil codes such as Article 6:99 of the Dutch Civil Code 21, and Art. VI.-4:103 of the DCFR. 22 The only difference is whether one of the defendants can escape liability by proving that his conduct was not the cause of the harm. According to the first clause of art. 10 TLL, one or more of the defendants can gain exoneration if and only if the true tortfeasor is determined. This strict restriction on defendants right to defense is claimed to be justified on the grounds of victim protection (Wang, 2013, p. 74). Relative to the victim, potential injurers are assumed to be much easier to prove who is the actual tortfeasor (ibid.). If defendants could obtain exoneration by proving his own innocence, it would be easier for all defendants to escape liability, leaving the victim uncompensated (ibid.). Case 2 concerns multiple sufficient causes or cumulative causation. In contrast to Case 1, it is already proven that each of D 1 s and D 2 s conduct (providing infected blood) is sufficient (1) BGB provides: If more than one person has caused damage by a jointly committed tort, then each of them is responsible for the damage. The same applies if it cannot be established which of several persons involved caused the damage by his act. The English translation of the BGB (German Civil Code) is available at < (accessed November 20, 2015). 21 Article 6:99 of the Dutch Civil Code provides: Where the damage is caused by two or more events, for each of which another person is liable, and it is ascertained that the damage originates from at least one of these events, then each of these liable persons is joint and several liable for that damage, unless a liable person proves that this specific damage is not caused by the event for which he himself is liable. The English translation of the Dutch Civil Code is available at < > (accessed November 20, 2015). 22 Art. VI.-4:103 DCFR provides: Where legally relevant damage may have been caused by any one or more of a number of occurrences for which different persons are accountable and it is established that the damage was caused by one of these occurrences but not which one, each person who is accountable for any of the occurrences is rebuttably presumed to have caused that damage. The outline edition of the Draft Common Frame of Reference (DCFR) is available at < (accessed November 20, 2015). 8

9 to cause the same damage (hepatitis C) independently. However, if the but for test is applied, neither D 1 s nor D 2 s conduct can be said to be the factual cause of the harm. It is argued to be extremely unfair if each of the Ds would escape liability just because there were other potential tortfeasors, especially when P were not contributorily negligent (Gilead, Green and Koch, 2013, p. 24). Case 2 can be resolved by applying art. 11 of the TLL, which reads: Where two or more persons commit torts respectively, causing the same harm, and each tort is sufficient to cause the entire harm, the tortfeasors shall be liable jointly and severally. Hence, it is evident that the Chinese approach to multiple sufficient causes is to hold all defendants liable in solidum. B. Parts of P s Harm Were Caused by Non-Tortious Factors Under tort law, the apportionment of damages usually occurs in three scenarios: (1) between multiple joint tortfeasors, (2) between the defendant and the plaintiff who is contributorily or comparatively negligent, or (3) between the tortfeasor and the non-tortious factor. As to the first scenario, art. 12 TLL provides that damages is apportioned according to each tortfeasor s seriousness of liability ( 责任大小 ) (evenly divided if it is hard to determine the seriousness of liability). According to the currently dominant scholarly opinion, the seriousness of liability is determined on the basis of both the degree of fault and causative potency ( 原因力 ) (Yang and Liang, 2006, 2009; Yang, 2009; Wang, 2013). Causative potency is defined as the contribution of each cause to the occurrence or expansion of an injury (ibid.). The seriousness of liability is primarily determined by comparing the degree of fault of each tortfeasor if they vary considerably in culpability (ibid.). If it is difficult to compare the degree of fault or in cases where strict liability applies, the seriousness of liability will be mainly based on the comparison of the causative potency of each cause (ibid.). When it comes to the second scenario, art. 26 TLL provides that damages that the defendant has to pay may be mitigated if the victim is also at fault as to the occurrence of harm, and art. 27 TLL gives total exoneration to the defendant if the harm is caused intentionally by the victim. The mitigation of damages is also argued to be determined by comparing both the degree of fault and causative potency (ibid.). Regarding the third scenario, if the non-tortious factor is outside the victim s own scope of influence such as a force majeure, the defendant will normally be exonerated from liability in accordance with art. 29 TLL. However, in the event that the non-tortious factor is within the victim s own scope of influence such as his underlying medical condition, vulnerabilities or predispositions, neither the new TLL nor the old GPCL has provided for any clear rules concerning how to deal with such non-tortious factors. Case 5 exactly concerns the third scenario. Case 5 is not associated with the problem of the but for test for factual causation, since it is already established, in light of the standard of proof, that D s conduct is the conditio sine qua non of (at least part of) the harm. In contrast to Cases 1 to 4, Case 5 is connected with causation in law, which deals with the question D should be held liable for what part of P s injury, i.e. whether and how damages should be apportioned between D and non-tortious factors. The third scenario represented by Case 5 is one of the most controversial issues concerning causal uncertainty in Chinese tort law. During the period , the applicable rules to the resolution of medical 9

10 malpractice claims were provided by Chapter 5 of the Regulation on the Handling of Medical Accidents (2002) 23 (hereinafter RHMA), in conjunction with the GPCL. The former provided detailed rules concerning the measure of different types of compensation for serious iatrogenic injury, while the later only stipulated a rough guide. Art. 49 para. 1 RHMA reads as follows: The following elements shall be taken into consideration in the determination of the specific sum of compensation for a medical accident: (1) The grade of the medical accident; (2) The degree of liability ( 责任程度 ) of the negligent medical act in the consequences of the injury caused by the medical accident; (3) The relationship between the injury caused by the medical accident and the state of the original illness. Sub-para. (2) and sub-para. (3) together provide a basis for apportioning damages between the negligent health care provider and the patient whose underlying medical condition also contributes to the occurrence or expansion of the final injury. It is evident that the patient s underlying medical condition cannot be regarded as contributory or comparative negligence on the part of the patient. Hence, the relationship between the negligent treatment and the underlying condition can only be understood by comparing the potency of the two causative factors. In practice, the degree of liability is determined by expert witnesses in the first place and reviewed and confirmed by the court finally. According to art. 36 of the Interim Measures for Medical Accident Technical Authentication (2002) 24 (hereinafter IMMATA) and Liu (2014), the degree of liability can be broken down into six categories as shown in Table 1. TABLE 1 DEGREES OF LIABILITY Code Theoretical Coefficient (%) Degree of Liability Recommended Compensation Range (%) A 0 No responsibility 0 B 10 Minor responsibility 1~20 C 25 Secondary responsibility 21~40 D 50 Equal responsibility 41~70 E 75 Major responsibility 71~90 F 100 Full responsibility 91~100 The RHMA s approach to Case 5 is argued to be applicable to other tort cases by analogy (Zhang and Ming, 2005). In practice, before 2014, this approach was also widely adopted in the trial of traffic accident cases. The publication of Guiding Case No. 24 (Rong v. Yongcheng Insurance Company) 25 by the Supreme People s Court, however, marked a turning point in the application of art. 49 para. 1 RHMA to other tort cases by analogy. In Rong v. Yongcheng Insurance Company, the ratio decidendi made clear by the appellate court is that if the victim of a traffic accident was not at fault, the influence of his vulnerabilities or predispositions on his final harm does not fall within any ground of justification that can mitigate the tortfeasor s compensatory liability. Accordingly, the appellate court held the defendant fully liable for all 23 Regulation on the Handling of Medical Accidents 医疗事故处理条例 (promulgated by the St. Council, Apr. 4, 2002, effective Sep. 1, 2002) 2002(15) ST. COUNCIL GAZ. 6, translated in (Lawinfochina). 24 Interim Measures for Medical Accident Technical Authentication 医疗事故技术鉴定暂行办法 (promulgated by the Ministry of Health, Jul. 19, 2002, effective Sep. 1, 2002) 2003(15) ST. COUNCIL GAZ Notice on the Publication of the Sixth Group of Guiding Cases 关于发布第六批指导性案例的通知 (promulgated by the Sup. People s Ct., Jan. 1, 2014, effective Jan. 1, 2014) (Chinalawinfo), translated by the author. 10

11 harm suffered by the victim. The authority of the Guiding Cases is high, for its selection and publication obviously indicates that the Supreme People s Court shares the same opinion as the appellate court. As a matter of fact, many local courts tend to refuse to adopt the RHMA s approach in traffic accident cases. 26 Therefore, the RHMA s approach is no longer applicable to traffic accident cases. Theoretically, Chapter 7 TLL has replaced Chapter 5 RHMA as far as medical malpractice liability is concerned. Since the TLL does not provides for any provision that is equivalent to art. 49 RHMA, it is fairly questionable whether it is still justifiable to continue to apply art. 49 RHMA to medical malpractice cases. For the time being, there is no evidence that courts have or will apply Guiding Case No. 24 to medical malpractice cases by analogy. Therefore, in practice, the RHMA s approach still persists at least as far as medical malpractice is concerned. C. The Hard Case and Lost Chances Case 3 and Case 4 actually belong the same category of cases where it is uncertain whether P is a victim of a tort and whether a single D, although acting tortuously (and therefore increasing the risk of harm), actually caused harm to anyone (Gilead, Green and Koch, 2013, p. 14). They are called hard cases because it is often difficult for the plaintiff to satisfy the standard of proof in order for causation in fact to be established. Case 4 They are hard cases also in the sense that no statute or judicial interpretation has ever provided a solution to these cases. It is interesting to see how courts handle cases similar to Case 3 and Case 4 in practice. 3.4 Empirical Findings In the end, a total of 592 medical cases closed for the period from a single grassroot level court (Gulou District Court) are included for this research. Three findings concerning causation are presented as follows: The court usually confirmed causation without giving much explanation. Of all the 280 cases, 246 reported how the causal link between the breach of duty and the damage was established. in about 38% (n=94) of the 246 cases, the court simply stated that there were no injuries but for the medical negligence without giving any argument. This more often than not occurred in cases where a direct infringement was involved. The court was most likely to apply a proportional liability to many cases where there was causal uncertainty. In almost 38% (n=95) of the 246 cases, the court identified that there was a certain degree of causality (even with a probability of less than 50%) between the negligence and the damage, such causality could not be excluded or there was indirect causality. The degree of causality would then be transferred into the degree of liability, against which the final awards are measured. 26 See e.g. Baohai Ying & Lanxiu Li v. Jiaxin Xue & PICC Zhijiang Branch 尹保海 李兰修与薛家新 中国人民财产保险股份有限公司枝江支公司机动车交通事故责任纠纷案, (Chinalawinfo) (Hubei Zhijiang County People s Ct. 2013) (China); Qinghao Sun v. Ansheng Tianping Insurance Company Jinan Branch 安盛天平财产保险股份有限公司济南中心支公司与孙钦浩等机动车交通事故责任纠纷案, (Chinalawinfo) (Shandong Jinan Intermediate People s Ct. 2014) (China). 11

12 Confronted by causal uncertainty, the court also switched the causal link to other kinds of injuries in some cases. Third, if there were no tangible personal injuries, the court might establish that there were a causal connection between the negligence and the loss of expectations or chances of being cured, survival or longer life (3.6%; n=9), pure emotional distress (8%; n=21) or pure medical costs (2%; n=4). Empirical evidence from the Gulou District Court demonstrates that the high probability standard of proof in China is almost totally circumvented by the court in malpractice lawsuits. On the one hand, it is found the court frequently adopts a proportional liability to many malpractice cases where there was causal uncertainty. In the court s opinion, a certain degree of causal link P* between medical malpractice and harm done even when the P* is as low as 10% - suffices to establish causation. Then the court would normally calculate pecuniary damages based on the product of the P* and the amount of pecuniary losses suffered by plaintiff-patients. On the other hand, in cases of misdiagnosis or breach of the duty of information disclosure, if the P* is relatively low, the court would switch the causal link between malpractice and physical harm to the causal connection between malpractice and the loss of chances of recovery, survival or longer life, or the infliction of pure emotional distress. There is evidence showing that this relaxation, or more appropriately, abandonment of the high probability standard of proof in medical malpractice cases is not unique to the Gulou District Court, but rather becomes a common practice nationwide (Liu, 2014). Judges seldom make clear in their judicial opinions why they choose to deviate from the high probability standard of proof. In an interview, however, one judge explains the rationale behind this deviation as follows 27 : Sometimes, expert testimony on causation is highly ambiguous, using wording such as there is a certain degree of causation or the possibility of the causal link cannot be excluded. We judges often feel confused by this wording and then we decide to use our discretion to interpret the testimony. If the defendant grossly breaches the duty of medical care and this breach of duty may sufficiently lead to iatrogenic injuries, we will rule that the causation is established. The word sufficiently implies a standard of proof relatively lower than the high probability standard. Otherwise it would be unfair on plaintiff-patients and their right to life and health could not be protected well [italics added]. It thus seems that the deviation is caused by both the ambiguity of expert testimony and by judges natural partiality to the weaker party plaintiff-patients. Nevertheless, this deviation is not founded on any legal basis. Whether this judicial practice persists and obtain its legal basis in the future is still unclear. 4 An Economic Justification for Proportional Liability Because of imperfect evidentiary information, it is not rare that there is considerable uncertainty about whether or not the injury is iatrogenic or to what extent the injury can be attributed to the health care provider s negligence. There are two main approaches to the 27 Interview with An Yuanyuan, Assistant Judge, Nanjing Intermediate People s Court, Jiangsu Province, China (Jul. 17, 2014). 12

13 determination of causation: the threshold probability criterion 28 and the proportional probability criterion. Under the threshold probability criterion (the all-or-nothing principle), injurers will not be held liable for harm done unless the probability of the causation between their actions and the accidents exceeds a given threshold (Shavell, 1987, p. 115; Scha fer and Ott, 2004, p. 209). No matter how high the threshold is set, two types of problems may arise (Shavell, 1987, p. 115). On the one hand, if the actual likelihood of causation were systematically below the threshold, then injurers would never be held liable for harm done. Under either strict liability or negligence, injurers would almost be completely undeterred. As a result, they would exercise too little care and engage too much in their activity. On the other hand, if the actual likelihood of causation were systematically above the threshold, then injurers would always be held liable under strict liability and would be much easier to be held liable under negligence. Consequently, they would take more care than is optimal or reduce their activity level too much under strict liability and would much more incentivized to take due care under negligence. Under the proportional probability criterion, however, parties will always be held liable for harm done unless the probability of causation is almost zero; and the damages will amount to the victim s losses times the probability of causation (Shavell, 1987, p. 115; Scha fer and Ott, 2004, p. 209). Under this proportional approach, the probability of causation equals the probability of the accident; therefore, the actual amount of damages paid by injurers is the same with the expected accident losses (Shavell, 1987, p. 116). This implies that the total accident costs facing injurers will remain unchanged before and after an accident and injurers will behave as they would in the absence of uncertainty over causation (ibid.). Nevertheless, the proportional approach is not immune from criticisms. One argument against this proportional approach is that it causes too many error costs than the threshold criterion does and it will be socially optimal if adjudicative error costs are minimized (Kaye, 1982). Shavell (1987, p. 117) refutes the argument by maintaining that it may be wrong to adopt the minimization of error costs as the social goal, because the recommending use of the threshold probability criterion will distort injurer s incentives to take efficient care. Another argument against proportional causation is that it incurs higher administrative costs than the threshold criterion for two reasons (ibid.): first, the volume of cases would be higher and a greater number of defendants would be involved; second, the chance of settlement would be lower and the cost of trial would be higher. Combining the advantages and disadvantages of proportional causation, Shavell argues that it should be applied where there is considerable uncertainty over causation (ibid.). Since causal uncertainty is prevalent and significant in medical malpractice cases, it may be socially optimal to apply the proportional probability criterion to the determination of the causation between breach of due medical care and iatrogenic injuries caused. 28 The prevalent criterion applied in common law countries is the preponderance of the evidence or balances of probabilities doctrine that sets a threshold of 50%. Civil law countries may take a higher threshold than 50%. 13

14 5. Conclusions By merely reading the TLL, Oliphant (2012) concludes: [b]y providing a general solution only to the problem of alternative or indeterminate defendants and restricting itself to an all-or-nothing approach, the new Chinese Tort Liability Law lags behind cutting-edge developments at national level, which have included the adoption of proportional liability in cases of alternative causes, both in the indeterminate-defendant scenario and where one of the possible risks emanates from the claimant s sphere. Of course, Oliphant was correct as far as only the TLL is concerned. However, if we look behind law on books and search in to law in action, we will find how active and flexible the role played by Chinese courts in applying legal rules is. Facing hard cases, Chinese courts systematically adopt proportional liability and the doctrine of loss of a chance, with a view to better protect patient s rights and interests. The proportional approach to hard cases can also be justified from the perspective of deterrence. In this respect, it is the Tort Liability Law rather than Chinese courts that lags behind cutting-edge development in other countries. However, it still remains to see how this judicial practice will develop in the future. 14

15 Appendix: Data Collection Methods The subject jurisdiction of Gulou District is situated in the center of Nanjing, the capital city of Jiangsu Province in the east of China. This project was designed to study retrospectively every medical malpractice case accepted and heard by the Gulou District Court from 2002 to Gulou District is largely an urban area, with a dense population of more than six thousand by 2011 (Gulou District Government, 2012) and an abundance of more than two hundred medical institutions, among which are twelve secondary or tertiary hospitals such as the Jiangsu Province Hospital, the Nanjing Drum Tower Hospital, the Nanjing Children s Hospital and the Zhongda Hospital of Southeast University (Qiu and Yao, 2012). Among all the eleven districts of Nanjing, Gulou District has the largest number and scale of medical institutions and almost half of all medical malpractice cases in Nanjing are accepted and heard by the Gulou District Court (Qiu and Yao, 2012). Before 2013, court decisions in Nanjing are mostly issued on the official website of the Nanjing Intermediate Court. I have collected a total of 162 medical malpractice cases from this website all of which were accepted by courts in Nanjing before the end of The results are displayed in Table 2. At least two problems deserve attention. First, as I mentioned before, court decisions prior to 2010 were not systematically issued on line, as evidenced by Table 1 where no case before 2010 has been found. Second, the subject jurisdiction of Gulou did not issue its court decisions consistently even after It follows that an analysis based on this on-line dataset has low level of internal validity. TABLE 2 - CASES ISSUED ON THE WEBSITE Year of Acceptance Number of Medical Malpractice Cases Total Gulou Court The Other Ten Trial Courts The Appellate Court Total Then, I diverted my attention to the Gulou District Court itself. I paid a visit to the Court in July 2014, explained my research project to judges there and eventually won their approval to have access to the court s internal case database. Cases prior to 2002 were rarely documented. I then searched every case involving medical injuries, medical malpractice, medical accidents or medical service contracts and downloaded judgments, in-court mediation papers and rulings in writing ranging from 2002 to I screened the downloaded case documents one by one with a view to excluding those that are not relevant to disputes over medical malpractice or that lack essential information. 29 In the end, a total of 592 medical cases are 29 Case documents lacking less important information such as the age, gender or places of residence of litigants are still kept for other analyses. 15