 | The 131st Annual Meeting (November 15-19, 2003) of APHA |
|
| The 131st Annual Meeting (November 15-19, 2003) of APHA |
William Fisher, PhD1, Steven M Banks, PhD2, and Bernice G Fernandes, BA2. (1) Center for Mental Health Services Research, University of Massachussetts Medical School, 55 Lake Avenue North, Worcester, MA 01655, 508-856-8711, bill.fisher@umassmed.edu, (2) Center for Mental Health Services Research, University of Massachusetts Medical School, Department of Psychiatry, 55 Lake Ave. North, Worcester, MA 01655
Persons residing in households often constitute the universe to be sampled in large scale epidemiologic and health service utilization surveys. These and other choices regarding inclusion and exclusion criteria in the sampling designs of such studies bring to the data they collect a set of "filters" which introduce explicit but also sometimes subtle limitations on generalizability of findings. These limitations often are ignored when prevalence, utilization rates, and correlations between risk factors are considered. This presentation illustrates these issues as they arise in the analysis of the National Comorbidity Survey (NCS). The NCS, which provides the most current data on the prevalence of, risk factors for and interrelationships among a range of psychiatric and substance abuse disorders, is based on a household sample of individuals age 15 - 54. Excluded from the sample were persons living in institutional settings, including hospitals, correctional settings, detoxification centers, residential programs and other settings in which persons with mental illness may reside. Significantly, some demographic groups, such as young African American males, may be at significant risk for exclusion because they have higher incarceration rates, often for offenses either directly or indirectly related to substance abuse. Some persons with severe mental illness also may reside in non-household settings, and thus be excluded from the sampling frame. The presentation will describe possibly flawed inferences that can be drawn from the NCS, discuss the policy significance of these issues, and identify other examples of household surveys in health services research in which similar issues arise.
Learning Objectives:
Keywords: Psychiatric Epidemiology, Criminal Justice
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.