Y Cao1, J S Krause, N DiPiro. 1. Department of Health Sciences and Research, College of Health Professions, Medical University of South Carolina, Charleston, SC 29425, USA. caoyu@musc.edu
Abstract
STUDY DESIGN: Cohort study. OBJECTIVES: First, to examine three sets of risk and protective factors for mortality after spinal cord injury (SCI), with an emphasis on health and secondary conditions. Second, to extend earlier work with several methodologic enhancements and addition of new predictors. SETTING: Twenty hospitals designated as SCI Model Systems (SCIMSs) of care in the United States. METHODS: Altogether, 8183 adults with traumatic SCI who received at least one follow-up evaluation between November 1995 and October 2006 from one of the SCIMSs were included in the study. There were 76,262 person-years and 1381 deaths at the end of June 2011. Mortality status determined by National Death Index and Social Security Death Index searches. Three successive sets of risk factors were evaluated with a logistic regression model on person-year observations to estimate the chance of dying in any given year. RESULTS: Several biographic and injury, socio-environmental and health factors were significantly related to the odds of mortality. A history of pneumonia or kidney calculus was associated with greater odds of mortality, whereas deep vein thrombosis was not. Poor general health, decline in health over the past year, hospitalization and a grade 3 or 4 pressure ulcer were also related to mortality. Consistent with a mediating effect, odds ratios declined with the addition of each successive set of factors. CONCLUSION: The relationship of biographic and injury characteristics with mortality after SCI is mediated by socio-environmental and health factors. Assessment of these variables enhances our ability to identify individuals at risk for excess mortality.
STUDY DESIGN: Cohort study. OBJECTIVES: First, to examine three sets of risk and protective factors for mortality after spinal cord injury (SCI), with an emphasis on health and secondary conditions. Second, to extend earlier work with several methodologic enhancements and addition of new predictors. SETTING: Twenty hospitals designated as SCI Model Systems (SCIMSs) of care in the United States. METHODS: Altogether, 8183 adults with traumatic SCI who received at least one follow-up evaluation between November 1995 and October 2006 from one of the SCIMSs were included in the study. There were 76,262 person-years and 1381 deaths at the end of June 2011. Mortality status determined by National Death Index and Social Security Death Index searches. Three successive sets of risk factors were evaluated with a logistic regression model on person-year observations to estimate the chance of dying in any given year. RESULTS: Several biographic and injury, socio-environmental and health factors were significantly related to the odds of mortality. A history of pneumonia or kidney calculus was associated with greater odds of mortality, whereas deep vein thrombosis was not. Poor general health, decline in health over the past year, hospitalization and a grade 3 or 4 pressure ulcer were also related to mortality. Consistent with a mediating effect, odds ratios declined with the addition of each successive set of factors. CONCLUSION: The relationship of biographic and injury characteristics with mortality after SCI is mediated by socio-environmental and health factors. Assessment of these variables enhances our ability to identify individuals at risk for excess mortality.
Authors: Amy L Strong; Annie C Bowles; Connor P MacCrimmon; Trivia P Frazier; Stephen J Lee; Xiying Wu; Adam J Katz; Barbara Gawronska-Kozak; Bruce A Bunnell; Jeffrey M Gimble Journal: Stem Cells Transl Med Date: 2015-04-21 Impact factor: 6.940