Heather Stuart1. 1. Dept. of Community Health and Epidemiology, Queen's University, Kingston (ON), Canada K7L 3N6. hh11@post.queensu.ca
Abstract
BACKGROUND: This secondary analysis describes the frequency of self-reported psychosocial risk factors in a geographically defined population of adolescents and quantifies the extent of multiple risks. Cluster analysis is used to develop three empirically distinct psychosocial risk clusters. METHODS: High school students in grades 9-13 from all seven public and two catholic high schools in the study area completed a class-administered survey. The analysis is based on 3540 surveys reflecting approximately a 71% response. Cumulative risk was calculated by summing the number of times students exceeded a pre-defined threshold on a series of global rating scales. Risk clusters were created using a non-hierarchical cluster analysis technique for binary data. Clusters were partially validated by examining differences in socio-demographic and health utilization patterns. Reliability was assessed by examining two, three, and four-group solutions across gender and grade strata. RESULTS: Multiple symptoms of emotional distress were reported by 37% of the sample, multiple stressors by 62% of the sample, and poly-drug use by 33%. In addition, three empirically distinct clusters were derived. Normals, 21% of the sample, did not report excessive stress or distress, and did not use substances. The Stressed (45%) reported excessive stress and distress predominantly related to schoolwork, parents, and facing problems. Virtually none used drugs. Substance Users (34%) reported excessive stress, distress, and high levels of substance use: smoking, drinking, and use of illicit drugs. Clusters were significantly different with respect to most socio-demographic factors, self-reported general health, and most aspects of health service utilization suggesting that they have some validity for targeting programs. CONCLUSIONS: This study highlights the importance of focusing on multi-morbidities and illustrates the use of cluster analysis to identify risk profiles that may be amenable to school-based health promotion and prevention programs.
BACKGROUND: This secondary analysis describes the frequency of self-reported psychosocial risk factors in a geographically defined population of adolescents and quantifies the extent of multiple risks. Cluster analysis is used to develop three empirically distinct psychosocial risk clusters. METHODS: High school students in grades 9-13 from all seven public and two catholic high schools in the study area completed a class-administered survey. The analysis is based on 3540 surveys reflecting approximately a 71% response. Cumulative risk was calculated by summing the number of times students exceeded a pre-defined threshold on a series of global rating scales. Risk clusters were created using a non-hierarchical cluster analysis technique for binary data. Clusters were partially validated by examining differences in socio-demographic and health utilization patterns. Reliability was assessed by examining two, three, and four-group solutions across gender and grade strata. RESULTS: Multiple symptoms of emotional distress were reported by 37% of the sample, multiple stressors by 62% of the sample, and poly-drug use by 33%. In addition, three empirically distinct clusters were derived. Normals, 21% of the sample, did not report excessive stress or distress, and did not use substances. The Stressed (45%) reported excessive stress and distress predominantly related to schoolwork, parents, and facing problems. Virtually none used drugs. Substance Users (34%) reported excessive stress, distress, and high levels of substance use: smoking, drinking, and use of illicit drugs. Clusters were significantly different with respect to most socio-demographic factors, self-reported general health, and most aspects of health service utilization suggesting that they have some validity for targeting programs. CONCLUSIONS: This study highlights the importance of focusing on multi-morbidities and illustrates the use of cluster analysis to identify risk profiles that may be amenable to school-based health promotion and prevention programs.
Authors: Chudley E Werch; Hui Bian; Joan M Carlson; Michele J Moore; Carlo C Diclemente; I-Chan Huang; Steven C Ames; Dennis Thombs; Robert M Weiler; Steven B Pokorny Journal: J Behav Med Date: 2010-07-27