The research proposed in this application is designed to capitalize on and further advance two exciting new developments in statistics, hierarchical (multi-level) statistical modeling, and meta-analysis, in order to address key issues in medical effectiveness, in health services, and in health policy research. The scope of problems addressed by these methods will include: modeling and quantifying variation in utilization rates and outcomes across study units (e.g., geographic areas, service areas, hospitals, or health care providers); and formulating improved random effects and meta-analysis approaches to combine diagnostic and therapeutic procedures. Hierarchical models may be used to provide better inferences for individual units by borrowing strength of information from the ensemble. They make it possible to account properly for different levels and sources of variation in complex datasets, and they often enable analyses of questions that cannot even be formulated by standard methods. This proposal brings together a network of statistical and health policy researchers from several Harvard departments, and two national advisory groups, one composed of leading statisticians in hierarchical models, meta- analysis, probability networks and health services research, and the other composed of senior subject-matter researchers from several PORT's. This research team will accomplish the transfer of new statistical technology to health services, outcomes and policy research, it will develop new statistical methodologies and software as needed, and it will disseminate its findings to the general community of researchers in these areas via a series of prototype analyses for specific problems.

Agency
National Institute of Health (NIH)
Institute
Agency for Healthcare Research and Quality (AHRQ)
Type
Research Project (R01)
Project #
5R01HS007118-03
Application #
2235945
Study Section
Special Emphasis Panel (HCT)
Project Start
1992-08-01
Project End
1996-07-31
Budget Start
1994-08-01
Budget End
1996-07-31
Support Year
3
Fiscal Year
1994
Total Cost
Indirect Cost
Name
Harvard University
Department
Administration
Type
Schools of Medicine
DUNS #
082359691
City
Boston
State
MA
Country
United States
Zip Code
02115
Ettner, S L; Christiansen, C L; Callahan, T L et al. (1997) How low birthweight and gestational age contribute to increased inpatient costs for multiple births. Inquiry 34:325-39
Gatsonis, C A (1995) Random-effects models for diagnostic accuracy data. Acad Radiol 2 Suppl 1:S14-21;discussion S57-67, S61-4
Rutter, C M; Gatsonis, C A (1995) Regression methods for meta-analysis of diagnostic test data. Acad Radiol 2 Suppl 1:S48-56;discussion S65-7, S70-1
Barsky, A J; Cleary, P D; Barnett, M C et al. (1994) The accuracy of symptom reporting by patients complaining of palpitations. Am J Med 97:214-21