Introduction to Ranking Models - Text Mining Techniques for Healthcare Provider Quality Determination - page 12

available to enable you to understand the data. MEPS currently has two major components for which
data are released: the Household Component and the Insurance Component. The Household Component
data are based on questionnaires to individual household members and their medical providers; the Insur-
ance Component estimates come from a survey of employers conducted to collect health insurance plan
information. These MEPS datasets have detailed cost information concerning all facets of healthcare
utilization, including physician visits, inpatient and outpatient care, medications, laboratory usage, eye
and dental care, and so on. The data for 2006 are available in December, 2008, with each subsequent
year added by December of each year. The data are available from the year 1996. Table 3 gives a list of
datasets available for the year 2005.
In the prescribed medicines file, there are a number of variables related to cost (Table 4). The RX
in the name relates to prescriptions. For the inpatient costs, 'RX' is replaced by 'OP', and by 'IP' for
inpatient costs. The '01' represents the year 2001; '05' is used for the 2005 data.
With such detailed cost information, we can examine the relationship between patient condition and
costs. Care must be taken when examining these datafiles, especially when merging cost information
from the different datasets.(Cerrito, 2008a)

nIs data

NIS data come only in ascii format, and the SAS coding needed to translate the datasets into standard
SAS format is available on the NIS web site. There is exactly one observation per patient, with no
longitudinal linkage between observations. There are two methods of coding patient conditions in the

Table 3. MEPS datasets for 2005

Dataset Name Description

HC-097 Full year consolidated file

HC-096 Medical conditions file

HC-095 Person round plan public use file

HC-094I Appendix to MEPS 2005 event files

HC-094H Home health file

HC-094G Office-based medical provider visits file

HC-094F Outpatient visits file

HC-094E Emergency room visits file

HC-094D Hospital inpatient stays file

HC-094C Other medical expenses

HC-094B Dental visits

HC-094A Prescribed medicines file

HC-091 Jobs file

HC-090 Full year population characteristics

HC-084 Population characteristics

HC-036BRR Replicates for calculating variances

HC-036 MEPS 1996-2005 pooled estimation linkage file