Soc Psychiatry Psychiatr Epidemiol (2008)

DOI 10.1007/s00127-008-0413-2

ORIGINAL PAPER

Ka Y. Liu Æ Eric Y.H. Chen Æ Ada S.Z. Cheung Æ Paul S.F. Yip

Psychiatric history modifies the gender ratio of suicide: an East and West comparison

Received: 17 May 2008 / Accepted: 2 July 2008 / Published online: 25 July 2008

j Abstract Objective Gender ratios of suicide rates differ greatly across countries. Victoria has a high male: female ratio in suicide that is typical in English-speaking and European countries, while in Hong Kong the low ratio is similar to other SouthEast Asian countries. This study investigates the effect of gender in the psychiatric and non-psychiatric populations to examine how psychiatric history may modulate the effect of gender in these two different communities. Method Review of coronial documentation of all suicide cases in 2000 in Hong Kong and Victoria, Australia. Log-linear model was used to test the gender-psychiatric history-location interaction. Results Gender difference in suicide was narrower in victims with psychiatric history than those without in both communities, albeit gender remained to have an effect even among those with a psychiatric history in Victoria. The impact of cultural- and gender-

specific factors is most apparent in suicide cases with no prior psychiatric history in Victoria: the male:female ratio was as high as 8:1 in this group of victims. Log-linear model results show that the gender-psychiatric history-location interaction was statistically significant. Conclusions The gender differentials in suicide rates in these two communities are mainly driven by gender’s effect in the population with no psychiatric history. Severe clinical conditions can override some, but not all, of gender’s effects in Victoria. Suicide prevention effort should target gender-specific factors to prevent suicides in men without history of psychiatric disorders. j Key words suicide – gender identity – Hong Kong – Australia – cross-cultural comparison

Introduction K.Y. Liu Institute of Social and Economic Research and Policy Columbia University New York (NY), USA E.Y.H. Chen Dept. of Psychiatry University of Hong Kong Pokfulam, Hong Kong A.S.Z. Cheung Austin Health Heidelberg (VIC), Australia Prof. P.S.F. Yip (&) HKJC Centre for Suicide Research and Prevention University of Hong Kong Pokfulam, Hong Kong Tel.: +852/2241-5017 Fax: +852/2549-7161 E-Mail: [email protected]

SPPE 413

P.S.F. Yip Social Work and Social Administration The University of Hong Kong Pokfulam, Hong Kong

The male:female ratios of suicide vary between countries. While this ratio is typically 3–4 to 1 in English-speaking and European countries, in some Asian countries it is often less than 2–1 [19]. China is the only country that reportedly has a higher suicide rate among women than men [3]. Several explanations have been put forward to account for this difference between the East and West. Phillips et al. [17] suggest that the high lethality of poisoning by pesticide, a suicide method commonly used rural women in China and other Asian countries, contributes to the high female suicide rates. A lack of resources for Asian women to deal with marital and family problems under a patriarchal culture may also have contributed to high female suicide rates [11, 15, 16]. Although these are likely explanations, research using comparable individual-level data collected from multiple locations is needed to test these hypotheses. Moreover, given the important role of psychiatric disorders in suicide, surprisingly little attention has

been given to how they may interact differently with gender in different populations. This study examines the interactions between gender and psychiatric disorders among those who have committed suicide in Hong Kong, China and Victoria, Australia. These two populations have roughly comparable population sizes, levels of economic developments and suicide rates. The population sizes in Victoria and Hong Kong were 4,785,000 and 6,712,000, respectively, in 2000. The suicide rate in Victoria in the same year was 12.1 per 100,000 and in Hong Kong it was 13.5 per 100,000. Yet Victoria has a high male:female ratio in suicide that is typical in English-speaking and European countries, while in Hong Kong the ratio is similar to those in South-East Asian countries. Our data were based on medical records of those who committed suicide in these two locations.

j Measures Details of lifetime psychiatric history for those who committed suicide were obtained from medical records collected by the police. They were classified into the following diagnostic categories: (1) cognitive (e.g. dementia); (2) schizophrenia; (3) other psychosis; (4) bipolar disorder; (5) depression; (6) substance/alcohol dependence; (7) anxiety; (8) personality disorders; (9) adjustment disorder; and (10) other diagnoses (i.e. developmental disorders, eating disorder, somatoform, dysthymic disorder, behavioural disorders, other infrequent disorders and disorders whose nature had not been specified). When the deceased had multiple diagnoses, they were ranked into primary, secondary and tertiary diagnoses in this order.

j Statistical analysis We used nested log linear modelling to examine the associations between location, gender and the presence of psychiatric disorders.

Results j Data Both Hong Kong and Victoria have a Coroner’s Court system of death registration. We reviewed the coronial documentation of all suicide cases in Hong Kong and Victoria in 2000. In 2000, there were 579 and 902 suicide cases in Victoria and Hong Kong, respectively. We used the same data collection protocol to review the two sets of coronial documentation to ensure data comparability. Excluding the 23 Victorian and 6 Hong Kong cases with missing information on psychiatric disorders, 556 Victorian and 896 Hong Kong cases were included in this analysis. In both locations, it is a routine police investigation procedure to examine public hospital records to determine if a person suspected to have died from suicide had a history of psychiatric disorder. However, the police have to rely on informants on whether the deceased had seen a private psychiatrist or other private health care professional for mental health problems. The demographic profiles of those committed suicide were then compared to the profiles of all psychiatric patients in the two locations. We obtained the gender breakdowns of psychiatric patients from the relevant government agencies. The gender breakdowns of those who were under public inpatient or outpatient psychiatric care in Hong Kong were provided by the Hong Kong Department of Health. Inpatient data were available for 2000 and outpatient data were available for 2003. Patients attending public psychiatric services were likely to have serious psychiatric disorders. Information on those who were attended by private psychiatrists was unavailable. Nonetheless, the number of psychiatric patients under private care is small in Hong Kong. It is estimated over 90% of all psychiatric-care related patient-days took place in public hospitals [9]. The gender breakdowns of those who attended state-funded specialist mental health services in Victoria in 2000–2001 were obtained from Victoria’s Administrative Services Database (RAPID). The specialist mental health services target people with serious mental disorders, and therefore the information should be comparable with the Department of Health data in Hong Kong. There is a greater diversity in the mental health care services for those with less severe mental health problems in Victoria compared to Hong Kong. For example, general practitioners and private psychiatrists provide a substantial part of the services [2]. Given the differences in service delivery and admission practice, we focused on the gender ratios of patients and it was not our intention to compare the absolute numbers of psychiatric patients in Hong Kong and Victoria.

j Psychiatric diagnosis While 58% of the suicide cases in Victoria ever had psychiatric records, only 31% of the cases in Hong Kong had such records. In Hong Kong, similar proportions among those who had psychiatric disorders had a primary diagnosis of depression (103/896 = 11.5%) and schizophrenia (100/896 = 11.2%). In Victoria, a high proportion of the cases had depression (200/556 = 34.0%), and a lower proportion had schizophrenia (35/556 = 6.3%). The third and fourth most common diagnoses in Hong Kong were adjustment disorders (22/896 = 2.5%) and other psychosis (13/896 = 1.5%); in Victoria they were other psychosis (25/556 = 4.5%) and bipolar disorder (25/556 = 4.5%).

j Gender and psychiatric history Figure 1 and Table 1 provide the gender breakdown of the suicide cases by the presence of lifetime psyMale

Female

50 40 Total Percentage

Methods

30 20 10 0 No diagnosis Victoria

Diagnosis

No diagnosis

Diagnosis

Hong Kong

Fig. 1 Total percentage of suicides by gender and life-time psychiatric history in Hong Kong and Victoria, Australia, 2000

Table 1 Number and percentage of suicides by gender and life-time psychiatric history and gender breakdown of psychiatric patients in Victoria and Hong Kong, 2000

Male N Suicides Victoria

No diagnosis With diagnosis With schizophrenia With depression Total Hong Kong No diagnosis With diagnosis With schizophrenia With depression Total Psychiatric patients Hong Kong Inpatients (2000) Outpatients (2003) Victoria Specialist mental health services (2000–2001) a

Female Percentage within gender

N

Male:female ratioa Percentage within gender

206 208 30 123 414 408 135 57 36 543

49.8 50.2 7.2 29.7 100.0 75.1 24.9 10.5 6.6 100.0

25 117 5 77 142 205 148 43 67 353

17.6 82.4 3.5 54.2 100.0 58.1 41.9 12.2 20..0 100.0

8.2 1.8 6.0 1.6 2.9 2.0 0.9 1.3 0.5 1.5

5,032 48,096 28,264

51.8 42.9 51.5

4,688 63,932 26,641

48.2 57.1 48.5

1.1 0.8 1.1

Number of males divided by number of females

chiatric history. Not surprisingly, the overall male: female ratio among suicide cases is higher in Victoria (2.9:1) than in Hong Kong (1.5:1). Closer examination reveals that the gender ratios were dependent on the presence of psychiatric disorder. The male:female ratios among those with a history of psychiatric disorder are 1.8:1 and 0.9:1 in Victoria and Hong Kong, respectively. Higher male:female ratios were found among those without such records in both locations. In Victoria, the male:female ratio among those who had no prior psychiatric history was extremely high (8.2:1). The same ratio was 2.0:1 in Hong Kong. Nested log linear models were used to test the 3-way interaction among gender, location and psychiatric history. The results show that the 3-way interaction was statistically significant (DG2 = 7.51, df = 1, P value <0.000). In other words, the effects of psychiatric history on the gender ratios of suicide seem to differ in Hong Kong and Victoria. As Fig. 1 shows, the gender gap in suicide rates was mainly driven by gender’s effect in the population with no psychiatric history. The gender ratios reported in Table 1 should be interpreted together with the gender breakdowns of psychiatric patients in the two populations as they may be merely reflecting differences in the gender compositions of psychiatric patients in Hong Kong and Victoria. However, the gender ratios among psychiatric patients were similar in the two populations. In Hong Kong the male:female ratios were 1.1:1 among all inpatients and 0.8:1 among out-patients, and it was 1.1:1 among those who attended the specialist mental health services in Victoria. In other words, while the gender ratio of those who had psychiatric history and committed suicide was comparable to the gender ratio of psychiatric patients in general in Hong Kong, there was an excess of male psychiatric patients died from suicide unexplained by

the gender compositions of psychiatric patients in Victoria. There were also differences in the gender ratios by diagnostic categories. Table 1 also shows the gender breakdowns for the two most common diagnoses in both populations: schizophrenia and depression. In both Hong Kong and Victoria, the male:female ratios are higher among the cases with schizophrenia than those with depression. The male:female ratio among those with depression in Hong Kong was only 0.5:1. Thus it is possible that the under-diagnosis of depression among men in Hong Kong had contributed to the low male:female ratio among the suicide cases with a history of psychiatric disorder in Hong Kong.

Discussion This study reported the prevalence of psychiatric history among suicides in Hong Kong and Victoria and the diagnostic differences, and examined the gender-psychiatric history interaction in these two populations. We have shown that the gender gap in suicide rates was mainly driven by gender’s effect in the population with no psychiatric history. The finding of a lower proportion of suicide cases with psychiatric history in Hong Kong than in Victoria is consistent with the findings from a previous study in Hong Kong [9] and figures from Japan [1]. The lower prevalence may be due to differences such as access to psychiatric services, diagnostic practices and the extent of under-diagnosis in Hong Kong and Victoria. There are important differences in the two mental health care systems: in Victoria, there is a greater diversity in the modes of service delivery, while in Hong Kong there is a strong reliance on inpatient care. Given the lack of choices, those who

suffered from psychiatric disorders in Hong Kong may be less willing to seek help from public health care than patients in Victoria. Furthermore, as awareness of psychiatric disorders are still lower in Hong Kong than in Victoria, under-diagnosis, especially of mood disorders, is a real possibility. To gauge the extent of under-diagnosis in the two populations, we need comparable epidemiological data on the prevalence of psychiatric disorders and service utilization rates in the two populations. Unfortunately, there has been no recent large-scale mental health study conducted in Hong Kong. The last study in the 1980s conducted in one area of Hong Kong finds that although the lifetime prevalence rates of psychiatric disorders, was close to those found in western societies, the prevalence rate of depression was lower [5]. The difficulties in comparing findings of epidemiological studies of psychiatric disorders are well known, however [6], it remains possible that psychiatric disorders have a less important role in suicide among Asian populations than in the West. For example, Phillips et al.’s [17] psychological autopsy study shows that the proportion of suicide victims with psychiatric disorder is lower in China than in Western countries. Further cross-country studies are needed to clarify the roles of service availability, utilization and suicide mortality rates in leading to the lower proportion of suicide victims with psychiatric history in Hong Kong. What this study adds is that psychiatric historygender interaction differed in the two populations. In both populations, psychiatric disorders moderate the effect of gender and bring closer the male and female differentials in suicide rates. We speculate that it is because severe psychiatric disorders have a strong impact on male and female patients alike and pose a strong risk of suicide regardless of gender. This is consistent with Qin et al.’s [18] finding that admission to psychiatric wards poses at least the same suicide risk in females as in males. Nonetheless, the higher male:female ratio of suicide of those with psychiatric history as compared to those without in Victoria suggests that non-clinical, gender-specific factors are likely to be operating even among those with a known psychiatric history. Consistent with our view that the cultural- and gender-specific factors have the biggest impact among non-psychiatric patients, the excess of male suicides is most apparent among those with no psychiatric history in Victoria with a male:female ratio as high as 8:1. The question that follows is whether the high male:female ratio among suicide cases with no psychiatric history is due to a higher prevalence of undiagnosed psychiatric disorders among men in Victoria. But as it has been often discussed in the literature, albeit of their low suicide rates in most Western countries, women are more likely than men to suffer affective disorders regardless of whether a diagnosis has been given [7]. According to the 1997

Mental Health and Wellbeing Survey [2], roughly equal proportions of men (17.8%) and women (18.0%) in Australia had at least one mental health disorder during the 12 months prior to interview. Findings from Chen’s study [5] suggest that it was also women who had a higher rate of psychiatric disorders in Hong Kong. Therefore, the prevalence of undiagnosed psychiatric disorders among men cannot explain the high male:female ratio among suicide victims with no psychiatric history in Victoria. As the gender difference in the prevalence of psychiatric disorders in the population cannot explain the gender differential in suicide rates, social and social-psychological factors have been suggested to account for the gender difference in suicide. As argued by many others, cultural-specific different norms and expectations for the two sexes lead to substantial differences in behaviours, which in turn lead to different health outcomes [14]. Such differences in gender roles may also explain the higher male suicide rates in Victoria than in Hong Kong. Several gender-related factors have been proposed to be at least partly responsible for higher suicide rates among men than women in Western countries in spite of women’s higher rates of psychiatric disorders. First, it has been suggested that men are less willing to seek help from professionals than women when they suffer from mental health problems [12]. While patient data from Victoria and Hong Kong suggest that men and women are equally likely to receive specialist psychiatric services, we cannot rule out the possibility that women are more likely than men to seek help from other health care professionals such as general practitioners for milder mental health problems and, therefore, subject to a lower suicide risk. As far coping strategies are concerned, it has been suggested that men are more prone to deal with their problems with alcohol and drug abuse, which increases their suicide risk [12, 13]. Moreover, men are believed to be more achievement-oriented than women, which make them more vulnerable to adversities such as unemployment and work-related problems [8]. Lastly, men may be less inhibited from violence and therefore more ready to use violent suicide methods, which make their suicide attempts more fatal [4]. Cross-cultural variations in these factors may account for the higher male:female ratios of suicide in Victoria than Hong Kong. For example, Phillips et al. [17] the high lethality of suicide by pesticide, a method commonly used by rural women, contributes to the low male:female ratio of suicide in mainland China. In Hong Kong, jumping from height and carbon monoxide poisoning by burning charcoal are the two most commonly used methods for both men and women [10]. The high lethality of these two methods may explain the lower male:female ratio of suicide in Hong Kong as compared to Victoria. We also speculate that the low prevalence of alcohol and drug use in Hong Kong may also have contributed to the lower

male:female ratio in suicide. Further cross-cultural comparison of risk factors, help-seeking behaviour and suicide method choice will further enhance our understanding of gender differences in suicide. Previous research typically relies on data at the aggregated level to explain cross-country differences in suicide and has limited our understanding on the individual-level processes that may have generated the differences. This study demonstrates the merits of using comparable individual-level data and our findings highlight the importance of looking at social factors such as gender roles in addition to clinical factors. In particular, our results suggest suicide prevention efforts should target gender-specific factors to prevent suicides in men without history of psychiatric disorders. This study has the following limitations. First, we only have information on the psychiatric records that were known to the police. No information on potential, undiagnosed psychiatric disorder was available. Absence of such records does not necessarily mean that the deceased had never had psychiatric disorders. Whilst the police had access to psychiatric services provided by the public health care system, information about psychiatric diagnoses made by private psychiatrists, general practitioners and other health care professionals varied, depending on the investigation. Although the presence of a psychiatric history is part of the routine police investigation for suspected suicide, in some rare cases the police may not have access to such information (e.g., family members were not aware of the diagnosis). Hence, the proportions of suicide cases with diagnosed psychiatric history may have been underestimated. We expect mild rather than severe psychiatric disorders are more likely to be underreported. However, such an underestimation should dampen our findings on a higher male:female ratio among non-psychiatric patients than patients only if the underestimation is more common among men than women. In contrast, it is women who are more likely than men to be attained by private psychiatrists, general practitioners and other health care professionals. Had we had a complete ascertainment on psychiatric history, we expect to find an even more striking male:female differential among those without a history of psychiatric disorders as compared to those without such history, i.e., a greater gender differential among those without psychiatric history while the gender ratio moves even closer to 1:1 among those with such history. Second, there may be differences in police investigation procedures and reporting in the two populations; but it seems unlikely that the differences are large enough to have generated the findings. Third, there are substantial differences in the mental health care systems in the two locations. Given such differences, we cannot draw any definitive conclusions on the absolute differences in the proportions of those with psychiatric history among those who committed

suicide in the two populations. Lastly, as the data consist of suicide cases only, no inference can be made about the differences in the relative risks of psychiatric disorders in the population. j Acknowledgments The authors would like to thank the Coroner’s courts in Hong Kong and Victoria for approving and assisting the data collection. We thank Will Lee, Michael Wong, Rhoda Cheung, Jade Au Brian Ip and Annette Graham for their assistance in data collection. We would also like to thank the Hong Kong Department of Health, HP Yuen from the University of Melbourne, Phil Barelli from the Information Development and Analysis Service Monitoring and Review Unit and Mr. Bill MacDonald from the Manager CAMHS and Youth Mental Health, Mental Health Branch of the Australian Government for their assistance in retrieving patient statistics in Hong Kong and Victoria.

References 1. Abe R, Shioiri T, Someya T (2007) Suicide in Japan. Psychiatr Serv 58(7):1013 2. Australian Bureau of Statistics (1998) Mental Health and Wellbeing: Profile of Adults, Australia 1997. Australian Government Publishing Service, Canberra 3. Bridges FS, Kunselman JC (2003) Rates of suicide in the world: 2002 update. N Am J Psychol 5(3):479–484 4. Canetto SS, Sakinofsky I (1998) The gender paradox in suicide. Suicide Life Threat Behav 28(1):1–23 5. Chen CN, Wong J, Lee N, Chan-Ho MW, Lau JT, Fung M (1993) The Shatin community mental health survey in Hong Kong. II. Major findings. Arch Gen Psychiatry 50(2):125–133 6. Cheng AT (2001) Case definition and culture: are people all the same? Br J Psychiatry 179:1–3 7. Culbertson FM (1997) Depression and gender. An international review. Am Psychol 52(1):25–31 8. Girard C (1993) Age, gender, and suicide: a cross-national analysis. Am Sociol Rev 58(4):553–574 9. Ho TP (2003) Psychiatric care of suicides in Hong Kong. J Affect Disord 76(1–3):137–142 10. Liu KY, Beautrais A, Caine E, Chan K, Chao A, Conwell Y et al (2007) Charcoal burning suicides in Hong Kong and urban Taiwan: an illustration of the impact of a novel suicide method on overall regional rates. J Epidemiol Commun Health 61(3):248–253 11. Mayer P, Ziaian T (2002) Suicide, gender, and age variations in India. Are women in indian society protected from suicide? Crisis 23(3):98–103 12. Moller-Leimkuhler AM (2003) The gender gap in suicide, premature death or: why are men so vulnerable? Eur Arch Psychiatry Clin Neurosci 253(1):1–8 13. Moscicki EK (1994) Gender differences in completed, attempted suicides. Ann Epidemiol 4(2):152–158 14. Nathanson C (1984) Sex differences in mortality. Ann Rev Sociol 10:191–213 15. Pearson V (1995) Goods on which one loses: women and mental health in China. Soc Sci Med 41(8):1159–1173 16. Pearson V, Liu M (2002) Ling’s death: an ethnography of a Chinese woman’s suicide. Suicide Life Threat Behav Winter 32(4):347–358 17. Phillips MR, Yang G, Zhang Y, Wang L, Ji H, Zhou M (2002) Risk factors for suicide in China: a national case-control psychological autopsy study. Lancet 360(9347):1728–1736 18. Qin P, Agerbo E, Westergard-Nielsen N, Eriksson T, Mortensen PB (2000) Gender differences in risk factors for suicide in Denmark. Br J Psychiatry 177(12):546–550 19. Yip PS, Callanan C, Yuen HP (2000) Urban/rural and gender differentials in suicide rates: east and west. J Affect Disord 57(1–3):99–106