Data and Measures Core

John Z. Ayanian, MD, MPP, principal investigator

The Data and Measures Core will provide the personnel, expertise, and
computational resources needed for effective utilization of the data to
be acquired and analyzed by the investigators in the Program Project.
This Core involves three main components. First, the Core will be
responsible for all data management and oversight activities including
dataset acquisition, preparation, integration, management, quality
control, security and archiving. Second, the Core will offer and
coordinate clinical expertise in measurement, specification of key
analytic variables, and interpretation across the projects.  Almost all
aims in the projects involve some clinical or health data. The explicit
clinical linkage will help ensure cross-project learning about fruitful
approaches and maximize consistency in approaches across projects.
Third, the Core will offer and coordinate expertise on statistical model
building and other data analysis methods across projects.  A
coordinated approach is particularly important given the common data
elements and interrelated research questions. The Data & Measures
Core is staffed by senior researchers with extensive experience with the
data and methods involved in the component projects. The Core’s primary
function will be an integrative one. This Core has the following
specific aims:

Aim 1
Acquire necessary data and create, integrate, and manage
data for projects. The Core is responsible for preparing data use
applications and obtaining and preparing the following major datasets to
be used in one or more P01 projects: Medicare Denominator file;
Medicare Part A and Part B claims for traditional Medicare (TM)
enrollees, Healthcare Effectiveness Data and Information Set (HEDIS®)
measures of quality and utilization for Medicare Advantage (MA) plans;
Consumer Assessment of Health Plans Survey (CAHPS) measures of patient
experiences for both MA and TM enrollees; Health and Retirement Study
(HRS) including restricted geographic identifiers and linked Medicare
claims; Medicare Current Beneficiary Survey (MCBS); Medical Expenditure
Panel Survey (MEPS); Community Tracking Survey (CTS) contextual
information; area level information from the Area Resource File (ARF)
and the US Census; and the Nationwide Inpatient Sample (NIS) and State
Inpatient Datasets (SID) from the Healthcare Cost and Utilization
Project (HCUP). The Core will monitor the progress of assembling these
key datasets, assist in the analysis, and advise each project on data
use and quality issues. The Core will construct program-wide datasets in
a common format on a regular basis, along with related documentation
and other tools to facilitate data use among the research projects.
These activities will promote integration of information across projects
and efficient use of large datasets. 

Aim 2
Coordinate and provide expert clinical input to all
projects. Senior clinical health services researchers will advise
research project leaders on the specification of clinical and health
status measures across the wide range of datasets to be employed in the
program project.  Expertise will be provided related to the use of
diagnostic and procedure codes and functional status data for risk
adjustment in all projects, specification of clinical quality and
utilization measures in Medicare claims data, and the specification of
clinically relevant subgroups of Medicare enrollees with specific health
conditions.

Aim 3
Coordinate and provide expert statistical input to all
projects.  Statisticians from the Core will advise project leaders on
both standard and innovative analytic techniques relevant to Medicare
and managed care data. Core statisticians will provide expertise in the
creation of composite quality measures, hierarchical modeling and causal
inference. A key focus of this effort will be to direct empirical
analyses and methodologic development related to creating composite
quality scores based on HEDIS and CAHPS data for both MA and TM
patients. The Core statisticians also have extensive experience in
developing geographically based hierarchical models that will be crucial
to the appropriate analyses of these complex datasets.