April 11, 2006
BIO656: MULTI-LEVEL STATISTICAL MODELS
SYLLABUS
Important links
·
Lectures
·
Homework
·
Lab assignments and related material
Main Texts
·
Goldstein (1995). Multilevel Statistical Models.
·
Hox (1995). Applied Multilevel Analysis.
Note: Both texts present a mix of
conceptual and quite technical material, all backed by excellent examples. Concentrate on the concepts and examples, but
be sure that you understand the connections between these and the formal
models. Unless you are a Biostatistics
graduate student, skip the technical derivations.
01. TUESDAY, MARCH 28th
Lecture (T. A. Louis)
·
Course overview and key concepts
·
Basic Multi-level modeling and
Shrinkage estimates
·
One-way ANOVA: Variance components
and standard errors
· Look
over the course web page
·
Goldstein: Chapters 0, 1, 2
·
Hox: 1
·
Download WinBUGS and begin reading
the manual
·
Singer J (1998). Using SAS PROC MIXED to Fit Multilevel Models,
Hierarchical Models and Individual Growth
·
Models. Journal of Educational
and Behavioral Statistics, 24(4): 323-335.
·
Diez-Roux A (2000).
Multilevel Analysis in Public Health Research. Ann. Rev. Public Health. 21:171-92.
Homework #1: (due Thursday, April 6th)
02. THURSDAY, MARCH 30th
Lecture (T. A. Louis)
·
Weighted means
·
More general linear MLMs
·
Examples of improving estimates
Readings
·
Jones AP,
Jørgensen SH (2003). The use of multilevel models for the prediction of road accident
outcomes. Accident Analysis and Prevention, 35: 59–69.
·
Goldstein on WinBugs
·
· Stanek EJ, Shetterley SS,
Allen LH, Pelto GH, Chavez A (1989). A cautionary note on the use of autoregressive models
in analysis of longitudinal data. Statistics in Medicine, 8: 1523-1528.
03. TUESDAY, APRIL 4th
Lecture (T. A. Louis)
·
Examples: Teacher Expectancy;
Diabetes control
Readings
·
Goldstein: 05 07
·
Hox: 2
·
Hofer et al. (1999). The
unreliability of individual physician “report cards” for assessing the costs
and quality of care of a chronic disease. JAMA, 281: 2098-2105.
·
Normand et al. (1996). Using
admission characteristics to predict short-term mortality from myocardial
infarction in elderly patients. Results from the Cooperative Cardiovascular
Project. JAMA, 125: 1322-1328.
·
Normand et al. (1997). Statistical
Methods for Profiling Providers of Medical Care: Issues and Applications. JASA,
92: 803-814.
04. THURSDAY. APRIL 6th
Lecture (T. A. Louis)
·
Non-linear
models: Conditional and Marginal
·
Logistic/binary
models
·
Crossover study
Homework #2: (due Tuesday, April 18th)
·
Heagerty PJ,
Zeger SL (1999). Marginalized Multi-level Models and Likelihood Inference. Statistical
Science, 15: 1–26. (advanced).
·
Palmer et al.
(1996). Results from a randomized controlled trial in 16 ambulatory care
group practices. Medical Care, 34, \#9: (Several articles).
05. TUESDAY, APRIL 11th
Lecture (M. Griswold)
· Non-linear
models, continued
· GEE
· Goldstein:
06 08
· Hox:
4
· Merlo J, Lynch JW, Yang
M, Lindström M, Östergren PO, Niels Kristian Rasmusen NK, Lennart Råstam L.
(2003). Effect of Neighborhood Social
Participation on Individual Use of Hormone Replacement Therapy and Antihypertensive
Medication: A Multilevel Analysis. AJE, 157: 774-783.
06. THURSDAY, APRIL 13th
Lecture (T. A. Louis and or
Mike Griswold)
· British
Social Attitudes Survey
· Bayesian
analysis and its relation to Multi-level Models
·
Liu J, Louis TA,
Pan W, Ma J, Collins A (2003). Methods for estimating and
interpreting provider-specific, standardized mortality ratios. Health
Services and Outcomes Research Methodology. 4: 135-149.
· Diehr P, Yanez D, Ash A, Hornbrook M, Lin DY (1999).
Methods for analyzing health care utilization costs. Ann. Rev. Public Health,
20: 125-144.
Homework #3: (due Tuesday,
April 25th).
07. TUESDAY, APRIL 18th
Lecture (T. A. Louis)
· Bayesian
Analysis, continued
· Goldstein:
10
· Hox:
4.1
· Affleck G, Tennen H, Keefe FJ, Lefebvre JC,
Kashikar-Zuck S, Wright K, Starr K, Caldwell DS (1999). Everyday life with
osteoarthritis or rheumatoid arthritis: independent effects of disease and
gender on daily pain, mood, and coping. Pain, 83: 601-609.
· Finch
et al. (2001). Substance use during pregnancy in the State of California USA.
Social Science and Medicine
08. THURSDAY, APRIL 20th
Lecture
(T. A. Louis)
· Ranking
· Discuss
Midterm
· Brief discussion of term
project
Readings
· O'Campo P (2003). Invited Commentary: Advancing Theory
and Methods for Multilevel Models of Residential Neighborhoods and
Health. AJE, 157: 9-13.
·
Peters ML,
Sorbi MJ, Kruise DA, Kerssens JJ, Verhaak PFM, Bensing JM (2000). Electronic diary assessment of pain, disability and
psychological adaptation in patients differing in duration of pain. Pain,
84:181-192.
·
Sampson RJ,
Raudenbush SW, Earls F (1997). Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy. Science, 277: 918-924.
Homework #4: (due
Tuesday, May 2nd).
09. TUESDAY, April 25th
Lecture (T. A. Louis)
· Ranking
continued
· Introduction
to Health Care Profiling
· Case
Studies: Normand, USRDS, Length of Stay
· Discuss
Term Project
· Goldstein:
11
· Hox:
3
· Diggle
et al. (1995). Statistical Issues in the analysis of disease risk near a point
source using individual or spatially aggregated data. Journal of Epidemiology and Community Health,
49: S20-S27.
· Shen W, Louis TA (2000). Triple-Goal estimates for
Disease Mapping. Statistics in Medicine, 19: 2295-2308.
·
Louis TA, Shen W (1999). Innovations in
Bayes and empirical Bayes methods: Estimating parameters, populations and
ranks. Statistics in Medicine,
18: 2493-2505
Homework:
Project
10. THURSDAY APRIL 27th
In-class EXAM:
· Open book; open notes
· 3 problems; approximately 30 minutes to read and
write down answers; 50 minutes to think
· Problems
will address issues such as:
-
Role of MLMs;
when and why to use, when/why not needed
-
Writing down a model; interpreting
coefficients
-
Shrinkage estimation
-
Linear and logit linear models
Basic computations
·
To be
assigned
Homework:
Prepare project proposal
11. TUESDAY, MAY 2nd
Lecture (M. Griswold)
·
Two-stage
modeling and joint modeling
· To be assigned
12. THURSDAY, MAY 4th
Lecture (F. Dominici)
· Applications
of Multi-level Models to Spatial Epidemiology
1. Spatial random
effects
2. Mapping
· Case
Studies: Pellegra, Scottish Lip cancer
Readings
·
Wakefield, J (2003). Sensitivity Analyses
for Ecological Regression. Biometrics 59, 9-17.
·
Wakefield, J (2003). Ecological Inference
in Epidemiology. Presentation at the 2003 Meeting of the International
Biometrics Society, Eastern North American Region.
Homework:
Project
13. TUESDAY, MAY 9th
Lecture (F. Dominici)
· The
National Mortality and Morbidity Air Pollution Study (NMMAPS) (Francesca
Dominici)
Readings
· Goldstein:
03 04
· Hox:
· Dominici F, McDermott A, Zeger SL, Samet JM (2003). National Maps of the Effects of Particulate Matter on Mortality: Exploring Geographical Variation. Environmental Health Perspectives, 111: 39-43
· Samet JM, Dominici F, Curriero F, Coursac I, Zeger SL 2002). Particulate Air Pollution and Mortality: Findings from 20 U.S. Cities. New England Journal of Medicine, 343: 1742-1757.
Homework:
Project
14. THURSDAY, MAY 11th
Lecture
(M. Griswold)
· Design: basic and advanced
· Design
example, tba
·
Rabe-Hesketh S,
Yang s, Pickles A (2001). Multilevel models for censored and latent responses. Statistical
Methods in Medical Research, 10: 409–427.
· Statistical Society Working Party on Performance Monitoring in the Public Services (2003). Performance Indicators: Good, Bad, and Ugly. Royal Statistical Society.
· Vaupel JW, Yashin AI (1985). Heterogeneity's ruses:
Some surprising effects of selection on
population dynamics. Am Statistician 39:
176-185.
Homework:
Project
15. TUESDAY, MAY 16th
Lecture
(T. A. Louis)
· Robustness
and efficiency
o
Working
o
Weighting trade-offs
o
Alternatives to the Gaussian
distribution for REs
· Goldstein:
09
· Hox:
5
· Dreams
Design Example
Homework:
Project
16. THURSDAY, MAY 18th
Lecture
(T. A. Louis)
· Missing
data
· If
time: Measurement error: Basic types, Attenuation/de-attenuation, shape change
· Coda
Homework:
Project due by midnight, Friday, May 19th.