K S Nicoll, M A McGee*, J S Callender

Royal Cornhill Hospital, Cornhill Road, Aberdeen;

*Health Services Research Unit, Medical School, University of Aberdeen, Foresterhill, Aberdeen.#

 

Abstract

Objectives: To review the changes in suicide and undetermined death rates in Grampian between 1991-1999 and to make comparisons

with a previous study on the same population. Design: All suicide and undetermined deaths in Grampian recorded by the General Register Office

for Scotland were included and linked with the psychiatric case records from Grampian Health Board. Results: The high suicide and undetermined

death rates in Grampian are accounted for by the excess of deaths in males. In males there has been a change in the method used to commit suicide with hanging now being the commonest method used. Firearm deaths have reduced dramatically since legislation was introduced in 1997 (RR = 0.21, 95% CI [0.05,0.91]). For those with previous psychiatric contact, a greater proportion were drug users than in 1974-1990 (OR = 3.75, 95% CI [2.7, 5.2]). A higher percentage of suicides have a history of more than one previous attempt at suicide than in 1974-1990. In Grampian there is a lower percentage of in patient suicide than the rest of Scotland. Conclusion; There have been changes in suicide trends in Grampian and these are similar to elsewhere in Scotland. Strategies to address this are discussed.

Key words – Suicide, undermined deaths, trends.

 

Introduction

In Scotland the rate of suicide in young men has continued to rise over the past two decades and is rising at a higher rate than the rest of the UK.1 We have looked at suicide trends in Grampian over a nine-year period. This follows on from a previous study looking at suicides in Grampian from 1974–19902 and attempts to explain why the rate of suicide in Grampian from 1989-1997 shows an increase.

in the suicide rate in Grampian over this period. This is in contrast to the national trend in Scotland, which shows the suicide rate to plateau out over the same period according to the Clinical Resource and Audit Group.3 Furthermore, between 1990 and 1997 a reduction in

suicide rates in England and Wales was noted.4 Mental illness is an important causal predictor of some suicides.1 However suicidal behaviour arises from a complex mixture of health, psychological, interpersonal, and social factors. The aim of this study was to review the changes in suicide and undetermined deaths rates in Grampian between 1991-1999 and discuss why this is not in keeping with national trends.

 

Methods

All deaths in Grampian by suicide (ICD 9, E950-959) and injury that was undetermined, whether accidentally or purposefully inflicted (ICD 9, E980-989), reported to the General Register Office for Scotland between January 1991 and December 1999 were examined. Data on age, sex and cause of death were obtained from Grampian Health Board.  Contacts between subjects and the psychiatric services in Grampian were identified through the records department at Royal Cornhill Hospital, Aberdeen. This has a computerised database containing records of all patient contacts with the psychiatric services in Grampian. A psychiatric contact was defined as being seen by a mental health professional of any discipline. Case notes of subjects with a psychiatric contact were reviewed. Details of type of contact at time of death, date of last contact, psychiatric diagnosis, and previous suicide attempts were abstracted. Diagnoses were assigned according to ICD-10 criteria.  To compare suicide rates by method we examined three three-year periods at the beginning, middle and end of the two consecutive suicide studies. Comparisons were made using two-sample t-tests, risk ratios and odds rations where appropriate.5 Suicide rates are calculated and plotted using three point moving averages to reveal trends.

 

Results

There were 752 deaths between 1991 and 1999 in Grampian in the groups described above. Of these, 551(73%) were recorded as suicides and 201(27%) were by injury that was undetermined. Suicide rates Figure 1 shows the suicide rate per 100,000 population by sex. The overall suicide rate peaks in 1997 and then plateaus out. Male suicide rates were cosiderably higher than female rates in all years and also peaked in 1997. Between 1991 and 1999 the suicide rate in males increased by from 18.2 per 100,000 person-years to 25.1 per 100,000 personyears.  The female rate has increased from 4.1 per 100,000 person-years to 8.1 per 100,000 person-year in the same period but has fluctuated more.  When looking at different age groups, young males, both with and without psychiatric contact became the group with the highest number of suicides. Males in the 25–34 age range had the highest number of suicides for time intervals 1994-1996 and 1997-1999 i.e. 51 out of 203(25%) subjects and 47 out of 217(22%) subjects, for the two time intervals respectively. The period 1991-1993 had the highest number of male suicides in the 35-44 and 45-54 age ranges, i.e. both groups had 33 subjects, the total for this time interval being 163.  

 

Method of suicide 

The suicide rates by method and sex for the periods 1974-76, 1988-90 and 1997-99 are described in Table I. For each period-sex combination the highest rate is highlighted in bold and underlined. Poisoning by solids or liquids is the commonest method of suicide for females in Grampian for each three-year period. Hanging is the commonest method of suicide for males in 1997-99. This has changed from the previous trends in Grampian when poisoning by solids or liquids was the most common method for males in 1974-1976 and poisoning from other gases (most commonly vehicle exhaust gas) was the commonest method of suicide in males in 1988-1990. The rate of suicide by hanging in males has more than doubled since 1988-90 and more than trebled since 1974–76.  Only 34(5%) of all the deaths between 1991-1999 involved the use of firearms and explosives. Of these deaths, 32 occurred between 1991-1997, prior to the changes to firearms legislation in 1997 made in response to the Dunblane incident, compared with 2 between 1998-1999 (RR = 0.21, 95% CI[0.05, 0.91]).  

 

Psychiatric contact 

Of the 752 deaths between 1991 and 1999, 310(41%) had previous psychiatric contact including 183(24%) who had psychiatric contact within a year of death. The commonest diagnoses in both sexes of those dying by suicide and undetermined deaths were affective disorder(34%), substance misuse(29%), neurotic disorders(17%), disorders of personality and behaviour(7%) and schizophrenia(6%).  In this study, 90(29%) of subjects with psychiatric contact had a substance misuse diagnosis compared with 36% in the 1974-1990 study. When substance misuse is divided into alcohol or drug related misuse 50% were drug related in 1991-1999 compared with 21% in the 1974-1990 study. Amongst all with a substance misuse diagnosis a greater proportion are drug users in 1991-1999 than in 1974-1990 (OR=3.75, 95% CI[2.7, 5.2]). Amongst men, opiate dependence was most common drug misuse category.  The time lapsed between the last contact with the psychiatric service and death is presented in Table II. There was a greater time lapse in deaths during the 1990s compared with 1974-1990 (mean difference = 77 days, p<0.001). The 1974-1990 figures only included subjects dying by suicide and not undetermined deaths.

 

Table III summarises the history of attempted suicide and hospital status at time of death. The profiles for the suicide and undetermined death groups are similar within the 1974-1990 (history x²=4.8 on 2d.f., p=0.09; status x²=1.8 on 2d.f., p=0.40) and 1991-1999 studies (history x²=0.7 on 2d.f., p=0.71;status x²=4.6 on2 d.f., p= 0.10).  Across all deaths, a history of more than one attempted suicide was more common in the 1991-1999 cohort than in the 1974-1990 cohort (OR=2.32, 95% CI[1.6, 3.4]) while the likelihood of being in psychiatric contact at the time of death was similar (OR=1.12, 95% CI[0.8, 1.5]).  However fewer were in-patients in 1991-1999 compared with 1974–1990 (OR=0.42, 95% CI[0.2, 0.7]).

 

Discussion

This paper outlines features in suicide specific to the Grampian region. The overall suicide rate peaks in 1997 and then plateaus out which is similar to the rest of Scotland except the peak occurred earlier in the decade.3 The high suicide and undetermined death rates in Grampian are accounted for by the excess of deaths in males. The male suicide rate has had more influence on the overall suicide rate. Suicide has been associated with areas of deprivation and unemployment.6,7 This does not explain the high rate in Grampian, which is one of the least deprived areas and had the lowest unemployment rate in Scotland in 1999.8,9  Our work uses routine data and case note review. Given the cumulative evidence that ‘missing’ suicides are more likely to be found among deaths labelled ‘undetermined’ than ‘accidental’, it has become customary to use the combined total of suicide and undetermined deaths as a guide to the ‘true’ total of likely or possible self-inflicted deaths.10 Eight subjects were excluded from having  psychiatric contact because they were recorded in the records department in Grampian as having been referred to psychiatric services but no case notes were located making it impossible to say whether they really had psychiatric contact.  Committing suicide depends to some extent on the knowledge and availability of methods. In males there has been a change over the past three decades in the method used to commit suicide. The introduction of catalytic converters in 1993 may explain the decreasing suicide rates from "other gas poisoning" for men since the early 1990s.  This is consistent with other studies, which suggest a link between suicide rate and the availability of effective means of committing suicide.10 The decrease in suicide rates due to poisoning by other gases in men was not paralleled by a decrease in the total suicide rate in Grampian. This is similar to elsewhere in Scotland where there has been no evidence of the influence of catalytic converters on overall suicide rates.12 The current high rate of hanging in males is consistent with other studies.11 It is a difficult area to address as we are unable to restrict use of this method.  Only a small number of cases (5%) of the total suicides and undetermined deaths in Grampian between 1991-1999 involved the use of firearms and explosives, which contrasts with the experience in the USA. In 1996 in the USA 62% of males and 42% of females who committed suicide did so by firarms alone.13 When the total of deaths by suicide and undetermined due to firearms and explosives are combined there is a reduction in the number of deaths by firearms since 1998 in Grampian. This could be secondary to the aftermath of the Dunblane incident in March 1996, and subsequent changes to the Firearms Act 1968 which came into effect in October 1997. In 1999 the number of firearm certificates and shotgun certificates on issue in Scotland were at their lowest recorded levels.14 In the USA suicide rates have decreased following the implementation of a variety of firearm control laws.15 This is evidence to suggest that the more restrictive the legislation, the lower the deaths by firearms.

 

The Department of Health introduced measures to restrict the sale of paracetamol and aspirin. This law came into effect September1998. It is too early to tell whether this is having an impact on suicide rates. It is well established that drug addicts and alcoholics are at an increased risk of suicide.11 In this study a higher proportion of those dying by suicide with previous psychiatric contact had a drug misuse diagnosis than in previous studies in Grampian. Substance misuse is one of the strongest predictors of suicide following non-fatal selfharm in young people.16 Since many drug users may not be in contact with psychiatric services then drug misuse could be a contributing factor to the increase in suicide rate of young males. Although many people would associate Scotland’s drug problem with the two cities of Glasgow and Edinburgh, it is striking that Aberdeen City has the third highest prevalence in Scotland behind Glasgow and Dundee.17 This finding confirms the need for enhancement of drug services. Between 1991-99, 24% of the total suicides in Grampian had contact with psychiatric services within a year of death. This is in keeping with national trends.18 Overall there has been a greater interval between last contact and death in those suicides with psychiatric contact in the 1990s compared with the period 1974-1990. This highlights the ongoing risk of suicide in subjects lost to psychiatric follow up. Double the percentage of people from the 1974-1990 study have history of more than one attempt at suicide.  This is obviously an important risk factor to consider when assessing deliberate self harm patients. At the time of death, less inpatients are dying by suicide in comparison to the 1974–1990 study. This is in keeping with trends in Scotland19 and could be partly related to a decline in hospital in-patient numbers. Only 41% of those subjects with previous psychiatric contact were having psychiatric treatment at the time of death, 35% being outpatients and 6% inpatients. This is lower than the percentage of in patient suicides in the rest of Scotland which has been noted to be 12%.18 This could be related to a number of factors including improved surveillance and care of in-patients. Suicide prevention is not solely the concern of mental health services. Recent studies have suggested psychiatrists cannot predict or prevent suicide with any reasonable level of accuracy at the level of the individual patients.20,21 Of the total suicides in Grampian, 76% had not been in contact with psychiatric services in the year before death. It is important that mental health services address any flaws in services and strengthen existing areas to facilitate suicide prevention. Improved liaison between mental health and substance misuse services and statutory and voluntary agencies may help. To reduce suicide we need a multi-agency approach and a broadly based strategy, incorporating health intervention, and targeting high risk groups. There are changes in suicide trends in Grampian. Continuous monitoring of this helps and will ensure preventative strategies are targeting the most appropriate demographic groups. Unfortunately this has not yet had an impact on the suicide rate in Grampian. Trends in suicide in Grampian are similar to elsewhere in Scotland.

 

ACKNOWLEDGEMENTS: The assistance of Pat Grant in accessing notes and setting up the suicide database is gratefully acknowledged. Disclaimer: Magnus McGee is employed by The Health Services Research Unit which is funded by the Chief Scientist Office of the Scottish Executive; however, the views expressed are those of the authors.

 

REFERENCES

1 Health in Scotland 2000. Armstrong EM. Ed. Stewart E. Chapter 3.Scotlands major health problem at the start of the 21st centuryPart3: Mental

Health:Suicide p32-33.

2 Ramayya A, Campbell M, Callender J S (1996) Death by suicide in Grampian 1974-1990.Health Bulletin, 54(1), 37-44.

3 Scottish Executive Health department. Clinical Resource and Audit Group (CRAG), July 1999 Clinical outcome indicators Report. Section A Updated Indicators. Pages 49-54.

4 McClure, G.M.G (2000) Changes in suicide in England and Wales, 1960- 1997.Br J Psychiatry, 176, 64-67.

5 Clayton D and Hills M. Statistical Models in Epidemiology. Oxford University Press,1993.

6 McLoone P (1996) Suicide and deprivation in Scotland.Br Med J.312: 543- 544.

7 Gunnell D et al (1999) Suicide and unemployment in young people. Br J Psychiatry 175, 263-270.

8 McLaren GL, Bain MRS, Deprivation and Health in Scotland: insights from NHS data. Chapter 2.Background Figure 1.Disribution of health board populations by carstairs deprivation category (1991 census) Edinburgh:ISD publications, 1998.

9 Scottish Economic Bulletin, No.58, 1999. 

10 Charlton J et al. Trends in suicide deaths in England and Wales. Population Trends 1992; 69: 10-16.

11 Kelly S, Bunting J (1998) Trends in suicide in England and Wales 1982-96. Population Trends 92; 29-41.

12 Kendell R E. (1998) Catalytic converters and prevention of suicides. The Lancet 352;1525.

13 Cutright P and Fernquist RM (2000) Firearms and Suicide : The American Experience, 1926-1996.Death Studies, 24, 705-719.

14 Scottish Executive press release: Firearm certificates Statistics Scotland 1999.

15 Lambert MT and Silva PS (1998) An update on the impact of gun control legislation on suicide. Psychiatric Quarterly. 69(2):127-34.

16 Hawton, K et al 1999 Suicide in young people. Study of174 cases, aged under 25 years, based on coroners’ and medical records. Br J Psychiatry175, 271-276.

17 Drug misuse statistics Scotland 2001.National Prevalence Study Page 23. 

18 Safety First: Five Year Report of the National Confidential Inquiry into Suicide and Homicide by people with Mental Illness. Report 2001.

19 The Mental Welfare Commission for Scotland: Annual Report 1999-2000.

20 Mortensen, P.B. (1999) Can suicide research lead to suicide prevention?Acta Psychiatrica Scandiavica,99,397-398.

21 Eagles J.M et al (2001) Role of psychiatrists in the prediction and prevention of suicide: a perspective from north-east Scotland. Br J Psychiatry,178, 494-496.