Body Mass and Suicide Risk - Hogrefe eContent

1 downloads 0 Views 67KB Size Report
1Wayne State University, Detroit, MI, USA, 2Richard Stockton College of New Jersey, Pomona, NJ, USA. Abstract. Body mass has been neglected as a possible ...
S. Stack & D. Lester: © 2007 Bo Crisis Hogrefe dy Mass 2007; &and Vol. Huber Suicide 28(1):46–47 Publishers Risk

Short Report

Body Mass and Suicide Risk Steven Stack1 and David Lester2 1

Wayne State University, Detroit, MI, USA, 2Richard Stockton College of New Jersey, Pomona, NJ, USA

Abstract. Body mass has been neglected as a possible predictor of suicidal behavior. The present investigation explores the association between body mass index and completed suicide. Data are from the National Mortality Followback Survey and refer to 373 suicides, and 518 deaths from motor vehicle accidents as controls. The results of a logistic regression analysis indicated that, for females, the greater the BMI the greater the suicide risk. However, this relationship did not hold up under controls for race and other covariates of BMI. For males, the BMI was not significantly related to suicide risk either at the bivariate or multivariate levels of analysis. The best predictors of suicide were living alone for females, and alcohol abuse for males. Keywords: body mass, suicide risk, predictor

There has been little research exploring the relationship of body size to suicidality. Lester (1981) found that nations with thinner children (that is, who were more ectomorphic) had higher suicide rates. Lester (1987) also found that the ectomorphy scores (height divided by the cube root of the weight) were associated with the choice of method for suicide. Those using suffocation were thinner (that is, more ectomorphic) than those using other methods. Thomas and Greenstreet (1973) found that medical students who later completed suicide were more ectomorphic in medical school. More recently, Magnusson, Gunnell, Tynelius, Smith, and Rasmussen (2005) studied Swedish male conscripts and found that there was a strong inverse association between height at the time of military service and later completed suicide – shorter men had a higher risk for suicide. Looking at the body mass index (BMI; height divided by the square root of the weight), Eaton, Lowry, Brener, Galuska, and Crosby (2005) found that students who perceived themselves as under- or over-weight had a higher risk of suicidal ideation and attempts than those who perceived themselves as normal in weight. BMI did not predict suicidality above and beyond this perception. The present study was designed to explore further the association between BMI and suicidality using a national sample of completed suicides.

Method The data were taken from the National Mortality Followback Survey (NMFS; US National Center for Health Statistics, 1990). The NMFS was based on a nationally representative sample of deaths in 1986. It refers to 18,733 adults, all over Crisis 2007; Vol. 28(1):46–47 DOI 10.1027/0227-5910.28.1.46

25 years of age, who died in the United States. The significant others of each decedent were interviewed. Complete data were available for 373 suicides (84 females and 289 males). Suicides were compared to deaths from motor vehicle accidents. Motor vehicle accidents were used as a control group since many natural causes of death (e.g., cancer) often involve weight loss, which affects body mass. The dependent variable is dichotomous wherein 1 = death by suicide and 0 = death by motor vehicle accident. Since the dependent variable is binary in form, logistic regression techniques are appropriate (Menard, 2002). The BMI was computed by dividing the decedent’s height in inches by the square root of their weight (in pounds); the greater the BMI the greater the tendency toward “thinness” or ectomorphism. Controls were introduced for alternative predictors of suicide (Lester, 2000; Stack, 2000a,b). Social integration was measured using two dichotomous variables: married vs. all others (0, 1), and living alone vs. all others (0, 1). Alcohol abuse was measured as a binary variable, i.e., the respondent was reported as drinking three or more drinks per day (0,1). Demographic controls included race where 1 = African American and 0 = all others, age (in years), and a series of binary variables for region of the country: Midwest (0, 1), South (0, 1), and West (0, 1), with the Northeast region serving as the benchmark category.

Results and Discussion The bivariate correlation between BMI and death by suicide was significant for females (r = 0.123, p < .05). However, controlling for the sociodemographic predictors, the © 2007 Hogrefe & Huber Publishers

S. Stack & D. Lester: Body Mass and Suicide Risk

BMI was no longer a significant predictor of female suicide (B = 0.3966, SE = 0.3437, Wald χ² = 1.11, p > .05). The variable with the strongest relationship to female suicide was living alone (B = 0.876, SE = 0.391, Wald χ² = 5.01, p < .05). From the odds ratio (OR = 2.40), women who lived alone were 2.4 times more likely to die from suicide than those not living alone. The Nagelkerke R2 statistic indicated that 11.5% of the variance was explained by the equation. For males, the bivariate correlation between BMI and suicide was not significant (r = 0.041, p > .05). The multivariate logistic regression analysis confirmed this null finding (B = 0.177, SE = 0.212, Wald χ² = 0.71, p > .05). The strongest predictor of male suicide was alcohol abuse (B = 0.416, SE = 0.176, Wald χ² = 5.59, p < .05). Male alcohol abusers were 1.51 times more likely to die from suicide than nonabusers. The Nagelkerke R2 statistic indicated that 3.9% of the variance was explained by the equation. As a check on these results, an index of ectomorphy (height divided by the cube root of the weight) was calculated instead of the BMI (height divided by the square root of the weight). Ectomorphy scores were also unrelated to suicide risk for both males and females at the multivariate level. These results, therefore, provide no support for the existence of an association between BMI and completed suicide, supporting the results of Eaton et al. (2005) for nonfatal suicidality.

References Eaton, D.K., Lowry, R., Brener, N.D., Galuska, D.A., & Crosby, A.E. (2005). Associations of body mass index and perceived weight with suicide ideation and suicide attempts among US high school students. Archives of Pediatric & Adolescent Medicine, 159, 513–519. Lester, D. (1981). Ectomorphy and suicide. Journal of Social Psychology, 113, 135–136.

© 2007 Hogrefe & Huber Publishers

47

Lester, D. (1987). Ectomorphy ratios of completed suicides. Perceptual & Motor Skills, 64, 86. Lester, D. (2000). Why people kill themselves. Springfield, IL: Thomas. Magnusson, P.K.E., Gunnell, D., Tynelius, P., Smith, G.D., & Rasmussen, F. (2005). Strong inverse association between height and suicide in a large cohort of Swedish men. American Journal of Psychiatry, 162, 1373–1375. Menard, S. (2002). Applied logistic regression analysis. Thousand Oaks, CA: Sage. Stack, S. (2000a). Suicide: A 15-year review of the sociological literature, part I: Cultural and economic factors. Suicide and Life-Threatening Behavior, 30, 145–162. Stack, S. (2000b). Suicide: A 15-year review of the sociological literature, part II: Modernization and social integration perspectives. Suicide and Life Threatening Behavior, 30, 163–176. Thomas, C., & Greenstreet, R. (1973). Psychobiological characteristics in youth as predictors of five disease states. Johns Hopkins Medical Journal, 132, 16–43. US National Center for Health Statistics. (1990). Public use datatape documentation: 1986 National Mortality Followback Survey. Hyattsville, MD: Author.

About the authors David Lester, PhD, is Professor of Psychology at The Richard Stockton College of New Jersey, Pomona, NJ, USA. Steven Stack is Professor of Psychiatry and Criminology at Wayne State University and Director of the Center for Suicide Research, Troy, MI, USA.

David Lester PhD Psychology Program The Richard Stockton College of New Jersey PO Box 195 Jimmie Leeds Road Pomona, NJ 08240-0195 USA E-mail [email protected]

Crisis 2007; Vol. 28(1):46–47

Suggest Documents