518 – Inference from Combined Data Sets
Approaches to Designing and Conducting Research on Hospitalization in View of the Reduced Samples in the 2008-2010 National Hospital Discharge Surveys
Margaret Jean Hall
National Center for Health Statistics
Maria F. Owings
National Center for Health Statistics
This paper describes challenges presented by reducing the sample size of the National Hospital Discharge Survey (NHDS) by half for the years from 2008-2010. NHDS, a nationally representative sample survey, was conducted annually from 1965 to 2010 by the National Center for Health Statistics, Centers for Disease Control and Prevention, and is the only source of trend data on hospitalizations during this period. Until 2007, data were gathered from about 400 hospitals and 360,000 discharges. In 2008-2010, data were gathered from about 200 hospitals and 166,000 discharges. Data from the half sample years are particularly important since they encompass the period when health care reform legislation passed, which is the baseline period to which any subsequent changes will be compared. This paper describes our findings on the effects of the half sample on data reliability for certain categories, and the implications of these findings for researchers interested in studying certain illnesses and patient groups.