The geography of self-injury: Spatial patterns in attempted and completed suicide

https://doi.org/10.1016/j.socscimed.2005.11.038Get rights and content

Abstract

This study uses hospital discharge data, death certificates and medical examiner data for New Jersey for 1999–2001 to investigate whether fatal and non-fatal self-injury exhibit similar geographic patterns. Findings show that the demographic characteristics of individuals committing fatal and non-fatal self-injury are quite different. Furthermore, attempted and completed suicides have a somewhat different geographical pattern. Municipality-level determinants of suicide and non-fatal self-injury were estimated in two separate models. While measures of isolation such as low population density and high proportions of households with only one member were predictive of completed suicides, non-fatal self-injury was related to unemployment and median age. Both types of self-injury were more common in municipalities which lost population between 1990 and 2000, and where divorce rates were high. Population-based suicide prevention efforts should be aided by the knowledge that although there are some similarities in the spatial pattern of completed and attempted suicides, there are some important differences in significant determinants.

Introduction

While suicide rates in New Jersey are among the lowest in the nation, suicide remains a significant cause of preventable mortality. Between 1999–2001 there were approximately 1770 suicides in the state. Furthermore, suicide follows a pronounced spatial pattern in New Jersey, in which rates are generally higher in the more rural counties, principally those located in the South and northwest of the state (Hempstead, 2002). This spatial pattern is strongest among white males, and mirrors a national geographic pattern in which suicide rates are highest in states with relatively low population density such as Wyoming and Alaska (Frankel & Taylor, 1992; Saunderson, Haynes, & Langford, 1998; Wilkinson & Israel, 1984; Zekeri & Wilkinson, 1995). Nationally, the age and geographical pattern of suicide among African–Americans is quite different, as rates for this population seem to be lower in rural areas (Willis, Coombs, Drentea, & Cockerham, 2003). The overall rural–urban suicide differential in English-speaking countries is well known and has been increasing in recent years (Singh & Siahpush, 2002). Far less is known about the spatial pattern of non-fatal self injury. Between 1999 and 2001, there were more than 12,000 non-fatal suicide acts in New Jersey which resulted in hospitalization. While New Jersey's suicide rates are low relative to those of the rest of the nation, rates of inpatient hospitalization for non-fatal self-injury are similar to the national average. As compared with completed suicides, those attempting suicide are more likely to be young, female and non-white.

Not all completed suicides are preceded by a suicide attempt, and the majority of attempters do not go on to commit suicide, as evidenced by the huge imbalance in annual numbers of fatal and non-fatal self-injury (National Institute of Mental Health (NIMH), 2004). In New Jersey, preliminary unpublished data from the 2003 National Violent Death Reporting System suggests that in 21% of completed suicides, there was a report of a prior attempt (Center for Health Statistics (CHS), 2005). Thus, the relationship between fatal and non-fatal self-injury is complex, and within the category of non-fatal self-injury there is much variation in seriousness of intent. The population of suicide attempters consists of relatively mild attempts, which are distinguishable from more severe attempts by use of less lethal methods such as wrist cutting and ingestion of medicines, as opposed to leaps from high places, suffocation and ingestion of toxic chemicals. Mild attempts are more common among females and adolescents and often result in negligible physical injuries. Since individuals who make mild attempts share few characteristics with completed suicides it has been argued that it may be more appropriate to refer to them as “parasuicide” rather than “attempted suicide”, the latter term which might be reserved for more serious attempts (Arensman & Kerkhof, 1996). While such a classification may be made in future analyses, for the present purposes suicidal behavior will be simply categorized into fatal and non-fatal groups.

Section snippets

Background

Research on suicidal behavior tends to focus either on demographic and social risk factors or underlying psychological processes. Mental illness, particularly depression, is widely acknowledged to be the single most important individual risk factor for suicide. In addition to depression, there are many other individual factors that appear to be significant, such as substance abuse, chronic pain, terminal illness or disability, domestic violence or a history of child abuse. Additionally,

Data and methods

Information on suicides comes from New Jersey death certificates and the multiple cause of death files, with some supplemental information obtained from medical examiner reports. Suicide is under-reported in all states, with some suicides misreported as unintentional or undetermined. This is particularly true for poisoning deaths. The extent of this under-reporting varies by state, and is not known for New Jersey. However, the proportion of poisoning deaths which are coded as undetermined is

Self-injury in New Jersey

As can be seen in Fig. 1, suicide rates are considerably lower in New Jersey than elsewhere although, in recent years, the gap has narrowed somewhat, primarily due to greater reduction in suicide rates in the rest of the nation. In 2001, rates rose slightly both in New Jersey as well as for the nation as a whole. Fig. 2 shows age-adjusted suicide rates by major mechanism. As can be seen, the major source of the difference in suicide rates between New Jersey and the nation as a whole is in the

Discussion

This study examined demographic and spatial characteristics of fatal and non-fatal suicidal behavior in New Jersey. There were two major objectives. One was to determine whether a rural–urban difference in suicide rates could be observed within a densely populated state such as New Jersey. Another goal was to see whether fatal and non-fatal self-injury had similar geographical distributions. With regard to the first objective, results suggest that even within a highly urbanized state such as

References (61)

  • A. Page et al.

    Suicide differentials in Australian males and females by various measures of socio-economic status, 1994–98

    Australian and New Zealand Journal of Public Health

    (2002)
  • T.E. Seeman

    Social ties and health: The benefits of social integration

    Annals of Epidemiology

    (1996)
  • P.S. Yip et al.

    Urban/rural and gender differentials in suicide rates: East and West

    Journal of Affective Disorders

    (2000)
  • E. Arensman et al.

    Classification of attempted suicide: A review of empirical studies, 1963–1993

    Suicide & Life-Threatening Behavior

    (1996)
  • R.D. Baller et al.

    Social integration, imitation, and the geographic patterning of suicide

    American Sociological Review

    (2002)
  • Bijou, Yang, & Lester. (2001). Learning from Durkheim and beyond: The economy and suicide. Suicide & Life-Threatening...
  • P.A. Boxer et al.

    Suicide and occupation: A review of the literature

    Journal of Occupational and Environmental Medicine

    (1995)
  • C. Branas et al.

    Rural–urban shifts in intentional firearm death: Different causes, same result

    American Journal of Public Health

    (2004)
  • Butterfield, F. (2005). Social Isolation, guns, and a ‘culture of suicide. New York Times, February 13, 2005, p....
  • T.M. Caldwell et al.

    Suicide and mental health in rural, remote and metropolitan areas in Australia

    Medical Journal of Australia

    (2004)
  • Center for Health Statistics. (2005). Unpublished data from New Jersey Violent Death Reporting System, 2003. New Jersey...
  • P. Congdon

    Suicide and parasuicide in London: A small-area study

    Urban Studies

    (1996)
  • B. Cox et al.

    Psychological dimensions associated with suicidal ideation and attempts in the National Comorbidity Survey

    Suicide & Life-Threatening Behavior

    (2004)
  • M.P. Cutchin et al.

    Scale, context, and causes of suicide in the United States

    Social Science Quarterly

    (1999)
  • M.J. Dudley et al.

    Suicide among young Australians: An interstate comparison of metropolitan and rural trends

    Medical Journal of Australia

    (1998)
  • E. Durkheim

    Suicide

    (1951)
  • M. Frankel et al.

    Suicide highest in wide-open spaces

    American Demographics

    (1992)
  • C. Gessert

    Rurality and suicide

    American Journal of Public Health

    (2003)
  • D. Gunnell et al.

    The relation between parasuicide, suicide, psychiatric admissions, and socioeconomic deprivation

    British Medical Journal

    (1995)
  • Hempstead, K. (2002). Suicide in New Jersey, 1999–2000. Center for Health Statistics:New Jersey Department of Health...
  • Cited by (60)

    • Child and adolescent suicidal and self-harm behavior

      2023, Encyclopedia of Child and Adolescent Health, First Edition
    • Demographic and socioeconomic correlates of suicide deaths and nonfatal self-injury related hospital visits: An analysis of counties in New York State

      2022, Psychiatry Research
      Citation Excerpt :

      Some evidence suggests that, like suicide, NFSI may be associated with indicators of social disintegration and socioeconomic disadvantage. For example, preliminary evidence suggests NFSI rates (defined using hospital records and discharge data) may be associated with regional elevations in unemployment, divorce, and poor childcare involvement (Hempstead, 2006; Philip and McCulloch, 1966). Like suicide, men and Whites are also more likely to be hospitalized for NFSI than women and racial minorities (Hanuscin et al., 2018); thus, NFSI rates may be higher in predominantly male, racially homogenous communities.

    View all citing articles on Scopus
    View full text