Innovative Methods for Modeling Longitudinal Medical Costs It is projected that health care costs per person would increase from $8,160 in 2009 to $13,100 in 2018, and that total health care costs will account for over 20% of the gross domestic product by 2018. Statistical analysis of medical cost data is becoming increasingly important with the heightened interests in containing the rising health care cost. Medical cost data are routinely collected in billing records of hospitals and claims of health insurance plans (e.g., Medicare, Medicaid, or commercial insurance). The wide availability of such data has motivated the development and application of the state-of-the-art statistical and econometric methods. With technological advances in automated data collection and management, medical costs are now often gathered at regular time intervals (e.g., daily or monthly), creating a longitudinal data pattern. The objective of this study is to develop and disseminate a number of models to analyze longitudinal medical costs data. There are five aims in this grant. First, we will expand the currently available econometric models of medical costs to longitudinal data and compare the performance of these models. Second, we will explore the use of more flexible functional forms of covariate specification in modeling longitudinal medical cost data. Third, we will extend the above models to jointly analyze medical costs and multiple health outcomes (e.g., survival, or quality of life), and study the effect of risk factors on them simultaneously. Fourth, we will apply hierarchical models to address the clustering effect in modeling longitudinal medical cost at different levels, e.g., health plans, families, and members. Finally, we will develop ready-to-use software to facilitate the practical application of methods developed from the proposed study. In addition to testing the performance of the proposed methods in simulation studies, these innovative methods will be applied to empirical case studies using three real-world databases: Clinical Data Repository (CDR) at the University of Virginia (UVA) Health System, Medical Expenditure Panel Survey (MEPS), and the SEER- Medicare databases. We expect the application of the proposed methods to these case studies will substantially advance our understanding of the influence of demographics, physician practice patterns, diseases, and health policies on the cost of medical care.

Public Health Relevance

Rising health care cost is a major concern for health policy makers. To better understand the factors associated with the growth in medical cost, it is important to study the longitudinal history of medical cost data. We propose to develop better methods to analyze longitudinal medical care costs data. To demonstrate the advantages of our proposed methods in clinical or policy decision making, we will apply these methods to a number of clinical- or policy-relevant case studies. We will also make programming codes of these methods available to other researchers who are interested in medical cost studies.

National Institute of Health (NIH)
Research Project (R01)
Project #
Application #
Study Section
Health Systems Research (HSR)
Program Officer
Henderson, Melford
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Northwestern University at Chicago
Public Health & Prev Medicine
Schools of Medicine
United States
Zip Code
Jagsi, Reshma; Bekelman, Justin E; Chen, Aileen et al. (2014) Considerations for observational research using large data sets in radiation oncology. Int J Radiat Oncol Biol Phys 90:11-24
Carter, Stacey C; Lipsitz, Stuart; Shih, Ya-Chen T et al. (2014) Population-based determinants of radical prostatectomy operative time. BJU Int 113:E112-8
Shih, Ya-Chen Tina; Xu, Ying; Dong, Wenli et al. (2014) First do no harm: population-based study shows non-evidence-based trastuzumab prescription may harm elderly women with breast cancer. Breast Cancer Res Treat 144:417-25
Zhao, Xingqiu; Deng, Shirong; Liu, Li et al. (2014) Sieve estimation in semiparametric modeling of longitudinal data with informative observation times. Biostatistics 15:140-53
Yu, Zhangsheng; Liu, Lei; Bravata, Dawn M et al. (2014) Joint model of recurrent events and a terminal event with time-varying coefficients. Biom J 56:183-97
Chen, Jinsong; Liu, Lei; Johnson, Bankole A et al. (2013) Penalized likelihood estimation for semiparametric mixed models, with?application to alcohol treatment research. Stat Med 32:335-46
Yu, Zhangsheng; Liu, Lei; Bravata, Dawn M et al. (2013) A semiparametric recurrent events model with time-varying coefficients. Stat Med 32:1016-26