Health-Care Utilization Due to Suicide Attempts Among Homeless Youth in New York State

Abstract

Suicide remains the leading cause of death among homeless youth. We assessed differences in health-care utilization between homeless and nonhomeless youth presenting to the emergency department or hospital after a suicide attempt. New York Statewide Inpatient and Emergency Department Databases (2009-2014) were used to identify homeless and nonhomeless youth aged 10-17 who utilized health-care services following a suicide attempt. To evaluate associations with homelessness, we used logistic regression models for use of violent means, intensive care unit utilization, log-transformed linear regression models for hospitalization cost, and negative binomial regression models for length of stay. All models adjusted for individual characteristics with a hospital random effect and year fixed effect. We identified 18,026 suicide attempts with health-care utilization rates of 347.2 (95% confidence interval (CI): 317.5, 377.0) and 67.3 (95% CI: 66.3, 68.3) per 100,000 person-years for homeless and nonhomeless youth, respectively. Length of stay for homeless youth was statistically longer than that for nonhomeless youth (incidence rate ratio = 1.53, 95% CI: 1.32, 1.77). All homeless youth who visited the emergency department after a suicide attempt were subsequently hospitalized. This could suggest a higher acuity upon presentation among homeless youth compared with nonhomeless youth. Interventions tailored to homeless youth should be developed.

Keywords: cost; health disparity; health-care utilization; homelessness; length of stay; suicide; youth.

Figure 1

Rates of suicide attempts by homeless versus nonhomeless youth according to month, New York, 2009–2014. The seasonal variance in the health-care utilization rate for suicide attempt among homeless and nonhomeless youth is depicted as monthly % difference from overall mean health-care utilization rate and was calculated as follows: 1) The overall mean rate of health-care utilization due to suicide attempt was estimated as the rate of emergency department visits and hospitalizations due to suicide attempts per 100,000 population during the study period of 2009–2014. 2) We calculated the rates by month. 3) We subtracted the overall rate from the monthly rate and divided the difference by the overall rate to obtain the % difference from the overall mean rate. The estimation was conducted separately for homeless (dashed line) and nonhomeless (solid line) populations. Global maximum occurred for both homeless and nonhomeless in May, followed by a global minimum in August. Local maximum occurred for both in October.