Alternative Options for State Level Estimates in the National Medical Expenditure Survey 1

This paper examines the issues associated with enhancing the sample design of the next cycle of the National Medical Expenditure Survey (NMES) Household Component, in order to allow for state level estimates of health care parameters. The NMES household survey was designed to produce estimates of the civilian non-institutionalized population for calendar year 1987. While small area estimation techniques can be used to produce model-based estimates with the current NMES-2 design, there are serious concern with respect to their accuracy and reliability. This evaluation reviews the coverage and precision requirements of other national surveys that produce state level estimates. Alternative options to enhance the capacity for producing state level estimates in NMES-3 are also examined.