Legal epidemiology of maternal deaths in Japan

an analysis on underreporting in vital statistics on Maternal & Child Health
paper presented before the delegation of U.S. Southern States' Legislators on Infant Mortality

Abstract

21 maternal deaths reported to have occurred during a period of eight years were analyzed to determine legal and epidemiological aspects of maternal deaths.
Uterine ruptures and postpartum bleeding were likely causes of death to develop dispute and inappropriate use of oxytocin might have been partially responsible for Japan's relatively high maternal mortality.
Female babies had higher survivability over male baby in fatally distressed situation of maternal death. Cases are also matched with vital statistics and at least four cases were identified to have occurred in prefectures where no maternal deaths were reported suggesting that attending doctors avoided indicating obstetrical causes in death certificate for fear of malpractice disputes.
This fact casts a serious doubt on the reliability of vital statistics and suggests that current data on maternal mortality is fatally underreported especially for disputable cases.

Introduction

Maternal mortality is a key indicator of a country's maternal and child health status. Japan achieved the world lowest infant mortality rate, however it is often pointed out that Japan's maternal mortality is not commensurate with it.
Japan's maternal mortality of 15.8 per 100,000 births in 1985 was higher than other developed countries such as France (12.0), Germany (10.7), the United States (7.8), the U.K. (7.3) and Canada (4.0)[1].
The reason of this inconsistency is equivocal. Yasukawa suspected the possibility of underreporting of maternal deaths in other countries, but failed to explain the fact that indirect obstetrical death accounts for more than half of the maternal mortality in other countries while it accounts for only 10% of Japan's maternal mortality[2].
Maternal mortality differs from infant mortality in that it is categorized by the cause of death instead of age. Age is an objective information and will be immune to biases, however maternal mortality may be biased by a variety of human factors.
In the first place, doctors who attended delivery may well hesitate in designating the cause of death directly related with the delivery because such deaths are highly likely to end up in malpractice disputes and the death certificate will be given to the survived family members as a nonconfidential document. Second, the true cause of death may not be so obvious unless the body undergoes autopsy and even attending doctors may not be able to tell the true cause with certainty. Third, there may be some biases when death certificates are collected and categorized by vital statisticians.
Although maternal mortality has steadily improved over the years, recent statistics shows a stagnation. The maternal mortality rate has been hovering around less than ten per 100,000 births. The 1993 figure dropped to 7.4 thanks mainly to a sharp decline of deaths due to postpartum bleeding.
The small number of maternal deaths makes it difficult to determine whether the changes are by chance or statistically significant. This supports a case-by-case approach rather than statistical analysis. Such case-by-case approach will be made possible by a centralized registry system like cancer registry.
Association of Japan Maternity Protection Doctors maintains a voluntary confidential reporting system to monitor maternal deaths[3], however only six cases were reported in 1991 and there will be a reporting bias in which disputable cases are less likely to be reported.
In view of these limitations, the author attempted to analyze the causes of individual cases obtained from published sources, many of which are disputed cases and hence deemed to have contributed to the high maternal mortality of Japan.

Materials and methods

Maternal deaths reported by mass media, academic case reports and court rulings were searched and retrieved by way of data base, newsletters of medical malpractice organizations and case books of court rulings.
Cases reported by mass media and ruled in the court are mostly disputed cases while cases reported as academic case report by doctors are mostly unavoidable and hence undisputed cases.
Searching for litigation news reported by mass media is indispensable because nearly half of the cases brought to court will end up in out-of-court settlement[4] and will not be included in the casebookof court rulings.

Results

21 identified cases are listed in chronological order of the date of maternal deaths in (Table) together with the data sources.
[chronological and geographic distribution]
Date of maternal death and prefectures seem are distributed reasonably reflecting the geographic distribution of the number of birth. Although the year 1987 and 1989 had four cases, the recent 1992 and 1993 had only one case respectively. The case which took place in 1992 involved no medical negligence. It seems that the number of disputable cases of maternal death is decreasing in recent years.
[matching with vital statistics data]
Vital statistics presents annual number of maternal deaths broken down by prefecture. The number of maternal deaths in a given prefecture in a given year is presented under the name of prefecture in parenthesis ( ). Four cases, notably Case 9,13, 19, 21, were from prefectures where no maternal death were reported in vital statistics. Those cases are denoted by asterisk * following the name of prefecture.
[status of dispute]
All cases except Case 15,17, 20 resulted in disputes. Case 1 constituted a criminal case in which the attending doctor was prosecuted and later convicted of one year and two months imprisonment with three years of parole[5]. All other 17 cases were civil disputes.
[age of pregnant women]
The age of pregnant women who lost lives ranges from 16 to 42, with the average of 31.6 years.
[causes of maternal deaths]
Uterine rupture was the major cause of death in seven cases (Case 9, 10,11,14,16,18,21), of which four cases were claimed by the survived family to have been caused by inappropriate use of oxytocin.
Excessive bleeding was the direct cause of death in five cases (Case 3, 6,13,19, 20), of which one case (Case 3) was explicitly determined to be due to inappropriate use of oxytocin by court ruling. Two cases were caused by premature separation of placenta.
[fate of fetuses/babies]
Two cases (Case 9, 12) were twin pregnancies and there were 23 fetuses involved, of which nine babies could survive. Of nine survived babies, eight were female and one was male. One female baby survived with severe brain damage (Case 18). Sex of aborted or dead fetuses were not available in all cases.
[damage claimed/awarded]
In 17 disputed cases, plaintiffs claimed an average of 86.6 million yen damage and in ten cases the defendants (hospitals and doctors) were required to pay an average of 45.8 million yen either by court rulings or through out-of-court settlement. The outcome of other cases are unknown. However there is no case in which defendants were exempted from liability.

Discussion

According to the vital statistics, there were 894 maternal deaths of direct obstetrical cause for a period of eight years from 1986 to 1993[6]. The 21 cases analyzed here are not a randomly sampled collection from the population and they can not be regarded as a representative of the entire maternal mortality. Because of this limitation, it is difficult to generalize the epidemiological findings of the sample to the entire population.
However the fact that at least one criminal case and 17 civil disputes derived from the relatively small number of maternal deaths signifies the highly litigious nature of maternal deaths and comparison with vital statistics will illustrate the cause of death which are likely to develop dispute.
Uterine rupture accounted for seven out of 21 cases (33.3%), which is considerably higher than the rate of uterine rupture of entire maternal deaths: 97 out of 894 (10.9%). Uterine rupture, for which some plaintiffs claim inappropriate use of oxytocin is responsible, is the likely cause of death for potential malpractice litigations.
Also the fact that hospitals and doctors were held liable in ten out of 17 disputed cases suggest there had been some sort of medical negligence. This finding lends support to the conclusion of an interim report of a study group on maternal mortality sponsored by Ministry of Health & Welfare saying that "half of the maternal deaths could have been avoidable"[7].
As for the sexual difference of survivability of babies, female babies showed a higher survivability over male baby (p=0.014). Although the sex of aborted or dead fetuses were unknown, the ratio of male to female of 14 aborted or dead fetuses will exceed the almost 2:1 ratio of entire fetal deaths in 1993. These findings clearly illustrate a large sexual difference in survivability in such a distressed situation as mother's death.
The most significant and potentially serious finding of this study is that four maternal deaths were confirmed not reported as maternal deaths in vital statistics. Vital statistics presents annual number of maternal deaths broken down by prefecture of dead mother's domicile and, thanks to the small number of maternal deaths, a growing number of prefectures especially those with less population have no maternal deaths.
However the author did identify four cases of obvious maternal deaths derived from prefectures where no maternal death had been reported. 　Case 9 was reported to have been issued a death certificate with a diagnosis of "cerebral bleeding" by the doctor and would have been classified as such in vital statistics. In Case 13, the mother died on 14 September 1989 after delivering a male baby on 5 August. There are 40 days interval between the delivery and the death. Since the definition of maternal death is "deaths caused by delivery within 42 days", this death should have been diagnosed as maternal death. Case 19 was probably diagnosed as "Multiple organ failure secondary to hemorrhagic shock" and Case 21 was also not included in vital statistics despite the incidence attracted wide media attention.
Unlike infant mortality, which can easily be classified by birth date, maternal mortality can not be classified precisely unless the attending doctor indicates that the death is related with delivery in death certificate. However the death certificate will be handed directly to the survived family members as non confidential document and any indication that the death is caused by delivery may prompt them to blame doctors for medical negligence.
In fact, the death certificate will always be presented to court as the first documentary evidence in most medical malpractice litigations. In view of these situation, it is a natural defensive action for doctors to avoid any indication which links the death with the delivery especially for potentially disputable cases.
The fact that at least four out of 21 maternal deaths are completely omitted from vital statistics gives us a grave warning in interpreting the data. Those four cases were identified simply because they happened to have occurred in prefectures where no maternal death was reported in a given year. Whether the rest 17 cases were reported appropriately will never be verified.
This study showed that the current statistics on maternal deaths are underreported and the true figures will never be available unless a strictly confidential compulsory reporting system like the one of U.K.[8]is instituted.

(Table)KNOWN CASES OF MATERNAL DEATHS 1986-93

No	date of death	prefecture	(#1)	mother's age	cause of death	appropriateness of oxytocin use	fate of fetus/baby	damage awarded/claimed(#2)
1	86-3-27	Tokyo	(17)	16	failed abortion		aborted (sex unknown)	unknown
2	86-4-19	Nagano	(5)	34	toxemia, premature separation		aborted (sex unknown)	---/61
3	86-11-8	Yamaguchi	(1)	23	bleeding (5830ml)	inappropriate use(ruled by court)	survived (female 2870g)	5.5/22
4	87-2-?	Fukuoka	(12)	36	pulmonary embolism (C section)		survived (male)	---/15
5	87-2-28	Saitama	(9)	19	ectopic pregnancy		aborted (sex unknown)	25 settled
6	87-12-?	Tokyo	(11)	29	bleeding (1814ml, C section)		survived (female)	40/100
7	87-12-29	Kyoto	(2)	28	Cephalopelvic disproportion	used	dead (sex unknown)	52.3/69
8	88-3-27	Tokyo	(11)	32	cerebral bleeding, toxemia		dead (sex unknown)	45/66
9	88-5-16	Oita*	(0)	31	uterine rupture (reported as CVA)	inappropriate use (claimed by plaintiff)	dead (twin, sex unknown)	---/82.2
10	88-9-3	Kanagawa	(11)	33	uterine rupture		aborted (sex unknown)	43/53.5
11	89-3-10	Kagoshima	(6)	40	uterine rupture		survived (female)	48/80.4
12	89-7-9	Shizuoka	(3)	29	toxemia,premature separation		dead (twin, sex unknown)	94/175
13	89-8-5	Hiroshima*	(0)	27	cervical laceration, bleeding	used (court ruled appropriate)	survived (male)	50/67
14	89-11-6	Hyogo	(9)	41	uterine rupture, bleeding (2600ml)	inappropriate use (claimed by plaintiff)	survived (female)	---/59
15	90-4-7	Tokyo	(11)	29	unknown		dead (sex unknown)	unknown
16	90-7-6	Fukuoka	(4)	33	uterine rupture	inappropriate use (claimed by plaintiff)	survived (female)	---/95
17	90-11-17	Kanagawa	(8)	37	eclampsia		survived (male)	not disputed
18	91-2-6	Shizuoka	(5)	35	uterine rupture		survived with brain damage (female)	---/282
19	91-6-24	Gunma*	(0)	32	bleeding		survived (female)	55/72
20	92-1-22	Ibaragi	(1)	42	bleeding (died at home)	not used	dead (female)	not disputed
21	93-5-28	Kyoto*	(0)	37	uterine rupture	so claimed by plaintiff	dead (sex unknown)	---/147.8

Legend and sources of (table)

Legend

#1: the number of maternal deaths reported in vital statistics in a given year. * denotes that the case occurred in a prefecture where no maternal death is reported in vital statistics.
#2: monetary unit of damage is million yen. "---" denotes the ruling is pending or unknown.

Sources

Case 1: Asahi Newspaper dated 5 August 1986, etc.
Case 2: Shinano-Mainichi Newspaper dated 16 May 1992.
Case 3: Hanrei-Times, No.824,1993:197.
Case 4: Nishi-Nihon Newspaper dated 30 October 1993.
Case 5: Akamatsu T. A case of maternal death caused by ectopic pregnancy. In: Medical Malpractice
Case Reports Vol.1. Nagoya:Medical Malpractice Information Center, 1991:4.
Case 6: Yui K. Medical Malpractiice. Tokyo: Futaba Co. Ltd., 1992:113-121.
Case 7: Yomiuri Newspaper dated 23 September 1994.
Case 8: Hanrei-Times, No.798,1992:230.
Case 9: Kyodo News dated 2 March 1990.
Case 10: Kyodo News dated 23 February 1995.
Case 11: Jiji Tsushin News dated 12 September 1995.
Case 12: Shizuoka Newspaper dated 29 September 1989.
Case 13: Mainichi Newspaper dated 1 December 1995.
Case 14: Yui K. Medical Malpractiice. Tokyo: Futaba Co. Ltd., 1992:86-112.
Case 15: Yomiuri Newspaper dated 9 April 1990.
Case 16: Nishi-Nihon Newspaper dated 24 October 1991, etc.
Case 17: Shinomiya H, et al. A case report of a sudden maternal death using head MRI. Japan Journal of Emergency Medicine Kanto 1991;12:108-110.
Case 18: Shizuoka Newspaper dated 11 January 1995.
Case 19: Mainichi Newspaper (Gunma edition) dated 21 June 1995.
Case 20: Mainichi Newspaper (Ibaragi edition) dated 24 January 1992.
Case 21: Mainichi Newspaper (Osaka edition) dated 13 March 1995.

References

[1].National Health Trend 1995. Tokyo:Health & Welfare Statistics Association, 1995:62.
[2].Yasukawa T, Nishida S, Hayashi K. An analysis of the secular trend of maternal mortality in Japan. Jpn J Public Health 1989;36:170-9.
[3].Honda H. On Maternal mortality monitoring system of Association of Japan Maternity Protection Doctors. Practice of Obstetrics & Gynecology1992;41:2053-9.
[4].Secretariat of Supreme Court. Handbook of civil litigation involving medical malpractice. Tokyo: Hoso-Kai, 1989:10.
[5]. Court ruling on 10 June 1987 by the first department of criminal case of Tokyo District Court.
[6].Ministry of Health & Welfare Minister's Secretariat Statistics and Information Department. Vital Statistics of Japan 1993. Tokyo: Health & Welfare Statistics Association, 1994.
[7]. Kobayashi N, et al. An interim report on maternal mortality. Ministry of Health & Welfare, June1991.
[8]. Department of Health and Social Services. Report on Confidential Enquiries into Maternal Deaths in the United Kingdom 1985-87. London. HM Stationary Office. 1991.
BACK
BACK to TOP
GP to CareManager.Net