Abstract
This study examined the extent to which posthospitalization change in connectedness with family, peers, and nonfamily adults predicted suicide attempts, severity of suicidal ideation, and depressive symptoms across a 12-month follow-up period among inpatient suicidal adolescents. Participants were 338 inpatient suicidal adolescents, ages 13 to 17, who were assessed at 3, 6, and 12 months posthospitalization. General liner models were fitted for depressive symptoms and suicidal ideation outcomes, and logistic regression was used for the dichotomous suicide attempt outcome. The moderating effects of gender and multiple attempt history were examined. Adolescents who reported greater improvements in peer connectedness were half as likely to attempt suicide during the 12-month period. Improved peer connectedness was also associated with less severe depressive symptoms for all adolescents and with less severe suicidal ideation for female individuals, but only at the 3-month assessment time point. Improved family connectedness was related to less severe depressive symptoms and suicidal ideation across the entire year; for suicidal ideation, this protective effect was limited to nonmultiple suicide attempters. Change in connectedness with nonfamily adults was not a significant predictor of any outcome when changes in family and peer connectedness were taken into account. These results pointing to improved posthospitalization connectedness being linked to improved outcomes following hospitalization have important treatment and prevention implications given inpatient suicidal adolescents’ vulnerability to suicidal behavior.
Acknowledgments
Support for this project was provided by a National Institute of Mental Health (NIMH) grant (R01 MH63881) and NIMH MidCareer Investigator Award (K24MH77705) to Cheryl King, Ph.D. We thank Sanjeev Venkataraman, M.D., and Paul Quinlan, M.D., for administrative and risk management support; and Anne Kramer, MSW, ACSW, Barbara Hanna, Ph.D., and Lois Weisse, R.N., for project management. We also thank our research staff, participating inpatient staff, and the adolescent participants and their parents and guardians.
Notes
Note: The Ns vary across time points between 301 and 337 for overall sample, 215 and 241 for females, 86 and 97 for males, 186 and 203 for non-multiple attempters, and between 115 and 134 for multiple attempters.
a Two-sample t-test assuming unequal variances (two-sided).
Note: For some variables n might be smaller due to missing data. MA = Multiple attempt.
a Paired-sample t-test (two-sided).
Note : N = 334.
a Two-way interaction between family connectedness change and time; The model also includes the variable of time necessary for the two-way interaction term.
N = 334; Two-way interaction between family connectedness change and multiple attempt status.
bThree-way interaction between peer connectedness change, time, and sex; The model also includes time, time x peer connectedness change, and time x sex necessary for the three-way interaction term.
Note: N = 252; OR = Odds Ratio; CI = Confidence Interval; p < .05 p < .01.