American Public Health Association
800 I Street, NW • Washington, DC 20001-3710
(202) 777-APHA • Fax: (202) 777-2534
comments@apha.org • http://www.apha.org

Statistics
Section Newsletter
Fall 2005

Message from the Chair

 
Marcia A. Testa, MPH, PhD Statistics Section Chair
Dear Statistics Section members,

As you now are all well aware, because of the devastating situation in New Orleans, the APHA 2005 Annual Meeting & Exposition has been moved. Our thoughts and prayers are with all Gulf Coast residents, especially our Section members, who have suffered great losses due to this terrible disaster. We know that rebuilding will be difficult, but as a Section we want to lend our support and help in any way possible.

I also want to make sure that everyone is still on board for the 133rd Annual Meeting and Exposition to be held in Philadelphia, Dec. 10 – 14, 2005. APHA is currently negotiating with several hotels with regard to its relocation of hotel reservations and will advise all attendees shortly. We appreciate everyone’s continued patience, understanding and support. To obtain updates on the new location please go to <www.apha.org>.

APHA is also working on updating this year’s meeting program to address the public health emergency across the Gulf Coast, relief efforts and rebuilding of communities and the public health infrastructure. As such, I believe that the Annual Meeting will provide an important opportunity for the public health community to collaborate, share across disciplines and construct long-term solutions to best address this tragedy.

I again want to send the Section’s congratulations to our most recent 2005 Statistics Section Award Winners. Martin Weinrich did a great job chairing the selection process. The 2005 winners will be honored on Tuesday, Dec. 13, 2005: 2:30 p.m.-4:00 p.m. during the Lowell Reed Lecture, Spiegelman and Statistics Section Awards Sessions.

2005 Award Winners

Academic: Carol K. Redmond
Government: G. David Williamson
Industry: Andy White

The revised schedule has all activities on the same calendar days as previously scheduled (e.g. Monday sessions will remain on Monday). Go to <http://www.apha.org/meetings/sessions.htm> for up-to-the-minute information; the online schedule reflects realtime changes.

In addition, the 2005 Officer Elects have been announced and at the end of the Annual Meeting will become:

Chair-Elect: G. David Williamson
Secretary-Elect: William Pan
Section Councilors: Amy Ferketich and Jim Leeper
Governing Councilor: Elizabeth Zell

I also want to introduce you to our current chair-elect and secretary-elect who will assume the positions of chair and secretary at the upcoming December 2005 meeting. You can become better acquainted with them by reading the following articles. I look forward to seeing everyone at the December meetings, and remember that the Annual Statistics Section Business Meeting and Reception will be held on Tuesday evening right after the Statistical Methods Sessions – please plan to come!


Sincerely,

Marcia A. Testa, MPH, PhD
Statistics Section Chair, 2005

133rd Annual Meeting - Statistics Section Program

3059.0: Monday, Dec. 12, 2005: 8:30 a.m.-10:00 a.m., Oral
Statistical and Modeling Techniques for Health Outcomes Research

Health outcomes research involves the collection and analysis of data that serves to provide evidence demonstrating the quality of medical care and public health programs. The assessment of patient-reported outcomes in clinical and health survey research has increased dramatically over the past several decades. Evaluations of symptoms, health status, quality of life and patient satisfaction are now primary and secondary outcomes in many clinical trials and observational studies. As indices and markers of therapeutic benefit and risk, these measures must be sensitive to differences that fall within a patient's operative range. Health care providers who must often prioritize resources across their members and among different conditions require a definitive metric. These evaluations typically use scales that have been developed to compare and contrast functioning relative to a given performance standard. To comprehensively evaluate the effectiveness of preventative programs and therapeutic interventions, investigators need to measure patient outcomes with the precision and accuracy that will allow for the detection of health improvements and health decrements that are important to the patient. The purpose of this session is to explore the various statistical, measurement and modeling techniques which can be used to evaluate the effectiveness of therapeutic interventions and prevention programs.


3137.0: Monday, Dec. 12, 2005: 10:30 a.m.-12:00 p.m., Oral
Measurement Issues: Race and Ethnicity

Public health surveys often collect data on racial and ethnic categories. Such information can be useful but can be difficult to obtain accurately and must be analyzed with care. This session will examine the statistical methods and implications for the use of these variables.
Learning Objectives:
1. Recognize and articulate the implications of differential reliability and validity for public health survey research.
2. Understand the use of scores for traditional minimal assumption Mantel-Haenszel tests and their extensions.
3. Learn about the latest research about variables that predict the single race category a multiple-race person will report as their primary race.
4. Understand the costs of treating 'Hispanic' as a quasi-racial category when studying health risks and behaviors.


3225.0: Monday, Dec. 12, 2005: 12:30 p.m.-2:00 p.m., Oral
Healthy People 2010: 2005 Midcourse Review Data Results and Methodological Issues



3225.1: Monday, Dec. 12, 2005: 12:30 p.m.-2:00 p.m., Oral
Design and Analysis of Community-Randomized Intervention Studies

Community- or group-randomized trials arise in many contexts, either because implementation logistics dictate such an approach, or because the total intervention effect is of interest when applied to a group of people. This session will treat both design and analytic features of group-randomized trials, and present a number of novel approaches of use to investigators and statisticians.
Learning Objectives:
1. Learn why constrained randomization can be useful in community-randomized trials, and how this can be implemented in a phased implementation trial.
2. Assess issues associated with spatial autocorrelation in the design and analysis of experiments.
3. Appreciate the importance of cost-effectiveness analysis of outreach interventions in community randomized trials.
4. Understand the two-stage approach to the analysis of community-randomized trials using generalized linear mixed models.


3317.0: Monday, Dec. 12, 2005: 2:30 p.m.-4:00 p.m., Oral
Innovations in Biostatistical Methods and Applications

The purpose of this session is to demonstrate new biostatistical and statistical techniques used in epidemiological, environmental health, health services and behavioral and social health science research. The relationship between the theoretical foundations of biostatistical methods and their recent applications are highlighted in this session. Learning Objectives:
1. Learn why bootstrap is necessary for multilevel modeling when the normality assumption is violated and/or the number of level-2 units is small.
2. Learn the form of a mixed effects model with linear splines with random knots.
3. Analyze various types of public health data under conditions of heterogeneity of variance.
4. Learn how to fit multidimensional mixture models to obtain cluster solutions.


3386.0: Monday, Dec. 12, 2005: 4:30 p.m.-6:00 p.m., Oral
Survey Research and Quantitative Methods for Health Care Services and Research

The role of quantitative methods is critical for evaluating the costs and effectiveness of health care services and programs. Many of these methods involve the creation, implementation, coordination and analysis of well-conducted surveys. This session will present state of the art methods and advances in survey research and analytical techniques necessary for providing and evaluating health care services.
Learning Objectives:
1. Acquire information and perspectives regarding the capacity of the Medical Expenditure Panel Survey to permit analyses of the U.S. health care system.
2. Articulate the strengths and weaknesses of the three different ways of determining the cost of illness as discussed in the presentation.
3. Learn the important differences between weighted and unweighted response rates.
4. Construct a statistical sampling strategy of US Census tracts to estimate countywide homeless population living in unsheltered conditions.


4062.0: Tuesday, Dec. 13, 2005: 8:30 a.m.-10:00 a.m., Oral
Evaluating Individual and Census Tract Characteristics on Birth Outcomes -- a Multi-level Approach

Learning Objective: To learn the use of multi-level analysis of contextual and individual characteristic which impact birth outcomes.


4094.0: Tuesday, Dec. 13, 2005: 12:30 p.m.-1:30 p.m., Poster
GIS Systems, Statistical Software, and Data Resources -- Posters I

Public health systems rely on timely and accurate data for optimal functioning. These posters provide information on important data sets, new ways to organize old data, and ways to incorporate geographic information into public health data bases and planning.
Learning Objectives:
1. Identify important features of large public health data sets.
2. Understand how geographic information can be incorporated into public health data sets.


4095.0: Tuesday, Dec. 13, 2005: 12:30 p.m.-1:30 p.m., Poster
Statistical Advances and Applications -- Posters II

A great variety of novel methodological approaches are applied to a great variety of questions in this poster session.


4259.0: Tuesday, Dec. 13, 2005: 2:30 p.m.-4:00 p.m., Oral
Lowell Reed Lecture, Spiegelman and Statistics Section Awards

The role of quantitative methods in public health and medical research is imperative to evaluating the effectiveness of new treatment programs and policies. The purpose of this session is to review and update the participator on examples of major areas of application of quantitative methods. Many different examples exist that can be used to demonstrate both the older and newer issues involved when applying measurement and statistical methodologies to a particular application such as: cancer prevention, disease diagnosis, public health screening, disease surveillance, drug development and safety monitoring, the conduct of clinical trials, evaluation methods for public health program evaluation, as well as genomics and genetics, environmental, maternal and child health, laboratory, epidemiological and health and social behavior research. Each year the Lowell Reed Lecture Series selects a distinguished individual that reviews a particular application of quantitative methodology to a public health research question. The session outlines the history, methods, applications and new advances for the quantitative methods described in the presentation.
Learning Objectives:
1. Describe the historical context of an application of quantitative methods to public health research.
2. Give examples of the measurement and statistical methods used in the application.
3. Evaluate the needs for future research and innovation in quantitative methods applications.
4. Chart out goals and objectives for further research in quantitiative methods applicable for solving public health problems.

Statistics Section Awards
Academia: Carol Redmond
Industry/Non-Profit: Andy White
Government: G. David Williamson
Spiegelman Award Winner: Rebecca Betensky

Lowell Reed Lecture – Statistics and Leadership: Our Legacy and Future, Jay H. Glasser, PhD, MS


4340.0: Tuesday, Dec. 13, 2005: 4:30 p.m.-6:00 p.m., Oral
Annual Statistical Methodology Session: Marginal Structural Models for Causal Inference

The purpose of this session is to introduce individuals in the fields of biostatistics, environmental health and epidemiology to the role of quantitative methods in public health research.
Learning Objectives:
1. Understand the utility of causal effects modeling.
2. Be able to identify situations that benefit most from causal modeling.
3. Learn how marginal structural models can be used to estimate causal effects.

436.0: Tuesday, Dec. 13, 2005: 6:30 p.m.-8:00 p.m., Business Meeting: Statistics Section Business Meeting and Social Hour
This very important session is open to all APHA members interested in the activities of the Statistics Section. New members are welcomed. The meeting will begin with a social hour so that members can network and learn about the activities of their colleagues and associates. This is an excellent opportunity to discover new contacts and make new friends. The social hour is followed by the annual Statistics Section business meeting during which Council representatives and officers will brief members about the Section's prior year's activities. Plans for the upcoming year will also be discussed. New officers will be inducted at this meeting. Please join us!

5079.0: Wednesday, Dec. 14, 2005: 8:30 a.m.-10:00 a.m., Oral
Factor Analysis: Developments and Applications

Dimensionality reduction techniques are critical to the analysis of many data sets that contain information on a large number of outcomes. In this session, factor analyses and related methods are proposed, described and evaluated
Learning Objectives:
1. Articulate how a k-fold likelihood cross-validation can be used to evaluate factor analysis models.
2. Apply differential item functioning techniques on scales and instruments used in health outcomes research.
3. Describe the advantages of a path model for investigating the influence of rural residence on health status.
4. Describe and distinguish between the various types of multiple indicator problems that complicate the evaluation of large-scale program initiatives.


5141.0: Wednesday, Dec. 14, 2005: 12:30 p.m.-2:00 p.m., Oral
Measurement Issues: Validity and Reliability

Self-reported data are invaluable to understanding many public health problems. These data, however, can be fraught with interpretational difficulties due to questions regarding their validity and reliability. In this session, methods are presented that can improve the collection and interpretation of personal data reported by individuals.
Learning Objectives:
1. Understand differential item functioning, item response theory, and the importance of establishing and employing valid measurement tools in epidemiological research.
2. Explain the effects of measurement error and violations of the true score model of reliability on test-retest estimates of the reliability of self-reported age of first drug use.
3. Discuss the importance of using psychometrically sound outcomes measures for clinical practice and research.
4. Test for the extent of differential reporting of similar levels of health among sub-groups of the population in any health surveys.
5. List the instruments commonly used in HIV/AIDS quality of life research.


5193.1: Wednesday, Dec. 14, 2005: 2:30 p.m.-4:00 p.m., Oral
Statistical Software and Science

Statisticians are engaged not only in finding simple solutions to complex problems, but also in constructing complex solutions. This session will present algorithms and their implemention for attacking difficult, yet intriguing, scientific problems.
Learning Objectives:
1. Analyze and evaluate different approaches to modeling of age-dependent seroprevalence changes.
2. Evaluate organ-matrix method as a cutting edge of technology to build systems.
3. Understand the algorithms behind imputation of genetic data.
4. Describe the nature of the relationships among ANOVA-type effect sizes and measures of association when there are two groups and when there are more than two groups.

Related Files:
133rdAnnualMeetingSTAT.doc

Meet the Chair-Elect

 
STAT Section Chair-Elect Larry Moulton
Get to know the next APHA Statistics Section Chair.

Larry Moulton is Professor of International Health at the Johns Hopkins Bloomberg School of Public Health, with a joint appointment in Biostatistics. He directs the Biostatistics Core of the CREATE consortium of group-randomized tuberculosis intervention studies, and is co-director of the Institute for Vaccine Safety. After receiving his BA in mathematics and Statistics at SUNY-Buffalo and his MS in biometry at the University of Texas, he served in Peace Corps as a statistician in the People's Republic of Benin. His doctoral work was in biostatistics at Johns Hopkins, and he was on the University of Michigan faculty before returning to Hopkins in 1991. Among other organizations, he serves as a consultant to the World Health Organization, the U.S. National Institutes of Health, the U.S. Food and Drug Administration, and the Survey Action Center of the Global Landmine Survey. He has served the Section in a variety of ways recently, including recruiting new primary and secondary members at the Section's exhibit hall table at the annual meetings and planning the program for the 133rd Annual Meeting. Larry will officially start his one-year term as Chair on Dec. 15, 2005.

Meet the Secretary-Elect

 
Secretary-Elect Brenda Gillespie
Brenda Gillespie is an assistant professor of biostatistics and associate director of University of Michigan's Center for Statistical Consultation and Research (CSCAR). She provides statistical collaboration and support for numerous research projects at the University of Michigan. She teaches biostatistics courses as well as CSCAR short courses in survival analysis, regression analysis, sample size calculation, generalized linear models, meta-analysis, and statistical ethics. Her major areas of expertise are clinical trials and survival analysis. She is planning the STAT Section Dinner at the 133rd Annual Meeting to be held Dec. 13 in Philadelphia. She can be contacted at <bgillesp@umich.edu> and at (734) 647-4609 (fax: (734) 647-2440).

STAT Section Election Results

We are pleased to announce the 2005 STAT Section election results. Officers in the "elect" positions will serve from Nov. 9, 2006-Nov. 7, 2007.

Chair-Elect:
G. David Williamson, MS, PhD, Centers for Disease Control and Prevention, Director, Agency Toxic Substances & Disease Registry (ATSDR), Office Assistant Administrator, Division of Health Studies;
PhD, Biostatistics, Emory University (1987);
Methods for disease surveillance, epidemiologic studies;
dxw2@cdc.gov

Secretary-Elect:
William Pan, MPH, MS, DrPH, Johns Hopkins University, International Health;
DrPH, Biostatistics, UNC-Chapel Hill (2003);
Multilevel and spatial models with GIS to examine environment and population health;
wpan@jhsph.edu

Section Councilors:
Amy Ferketich, MA, MAS, PhD, Ohio State University, Epidemiology & Biostatistics;
PhD, Public Health, Ohio State University (2000);
Lifestyle, nutritional and other risks for cancer;
aferketich@sph.osu.edu

Jim Leeper, MS, PhD, University of Alabama, Community and Rural Medicine;
PhD, Biostatistics, Iowa (1977);
Infant mortality, teen suicide, children's health, missing data in longitudinal analysis and spatial/temporal analysis;
jleeper@cchs.ua.edu

Governing Councilor:
Elizabeth Zell, MS (Stat), Centers for Disease Control and Prevention;
Infectious Disease Surveillance; ezr1@cdc.gov

Statisticians and De-Identifying Protected Health Information for HIPAA

The Health Insurance Portability and Accountability Act of 1996 (HIPAA) is mostly known among health care providers and researchers for repercussions of its Privacy Rule on work practices. HIPAA has had great impact on public health research due to restrictions on access to databases with any of the 18 Protected Health Information (PHI) variables. Most statisticians are aware of HIPAA because of required human subjects training and changes in database sharing, data use agreements, and security policies.

However, an important component of the HIPAA Privacy Rule is:
The second way to de-identify PHI is to have a qualified statistician determine, using generally accepted statistical and scientific principles and methods, that the risk is very small that the information could be used, alone or in combination with other reasonably available information, by the anticipated recipient to identify the subject of the information. The qualified statistician must document the methods and results of the analysis that justify such a determination (NIH, 2004).

Moreover, the de-identifier must not “disclose the key or other mechanism that would have enabled the information to be re-identified”; this includes not divulging pseudo-random number algorithms or seed values.

The two phrases in the quoted text, “very small [risk]”and “qualified statistician”, are examined further.

The HIPAA Rule does not quantify “very small,” but the Subcommittee on Disclosure Limitation Methodology of the Federal Committee on Statistical Methodology has provided an oft-cited 1994 working paper on the issue (FCSM, 1994). The working paper cites the National Center for Education Statistics Standard for Maintaining Confidentiality (IV-01-91): ‘In reporting on surveys and preparing public-use data tapes, the goal is to have an acceptably low probability of identifying individual respondents.’ The standard recognizes that it is not possible to reduce this probability to zero.” Although a specific probability is not stipulated, the standard does state that “cells be based on at least three unweighted observations and subsequent tabulations (such as crosstabulations) must not provide additional information which would disclose individual identities.”

The U.S. Office of Personnel Management defines two general schedule positions, Statistician (GS-1530) and Mathematical Statistician (GS-1529), with different qualification standards, which may add to the confusion. The educational requirements of the two positions each dictate as few as six semester hours (about two courses) in statistics with enough hours in mathematics (nine for statistician and 18 for mathematical statistician; details at OPM links below).

However, upon closer examination, the actual text of the Rule itself (Section 164.514(b)(1)) does not mention “statistician” but instead says “a person with appropriate knowledge of and experience with generally accepted statistical and scientific principles and methods for rendering information not individually identifiable.” The more widely disseminated interpretations use the phrase “qualified statistician” defined as “a person with appropriate knowledge of and experience with generally accepted statistical and scientific principles and methods for rendering information not individually identifiable” (NIH, 2005).

Finally it is important to note that, if the de-identification was not properly performed or the probability was not small enough, some believe statisticians may be subject to prosecution (Dressler, 2005), since individuals in addition to covered entities have been charged and sentenced for infractions. However, a 2004 Seattle case of individual prosecution involved malicious identity theft. Furthmore, this summer a Department of Justice memorandum seems to indicate HIPAA only applies to covered entities not individuals (Pear, 2005).

Thus, public health statisticians who otherwise think of themselves as “qualified statisticians” may want to re-evaluate their training and experience with regard to the HIPAA Privacy Rule and their own risk exposure until the issue is resolved definitively.



References:

Dressler LG: Human specimens, cancer research and drug development: How science policy can promote progress and protect research participants. National Cancer Policy Board for the Institute of Medicine and the National Research Council, March 2005, p39.

HIPAA Privacy Rule and Its Impacts on Research, NIH website http://privacyruleandresearch.nih.gov/healthservicesprivacy.asp,
Updated 04 Aug 2004, accessed 16 July 2005.

Health Services Research and the HIPAA Privacy Rule, NIH website, NIH Publication Number 05-5308
http://privacyruleandresearch.nih.gov/pdf/HealthServicesResearchHIPAAPrivacyRule.pdf
Updated May 2005, accessed 16 July 2005.

Federal Committee on Statistical Methodology, Report on Statistical Disclosure Limitation Methodology, Statistical Policy Working Paper 22, Washington, DC: Office of Management and Budget, 1994.
http://www.fcsm.gov/working-papers/wp22.html
Updated May 1994, accessed 16 July 2005.
http://www.fcsm.gov/committees/cdac/

Office of Personnel Management
http://www.opm.gov/qualifications/SEC-IV/B/GS1500/1530.HTM
http://www.opm.gov/qualifications/SEC-IV/B/GS1500/1529.HTM
Accessed 16 July 2005.

Pear R: Ruling Limits Prosecutions of People Who
Violate Law on Privacy of Medical Records, NY Times, 07 June 2005, A16.


2005 International Conference on Health Policy Research Oct. 28-30

 
The 2005 International Conference on Health Policy Research (ICHPR), which will be held Friday, Oct. 28 - Sunday, Oct. 30, 2005, at the Boston Park Plaza Hotel, is open for registration. ICHPR focuses on methodological issues in health services and outcomes research. It has been held every other year since 1995, sponsored by the Health Policy Statistics Section of the American Statistical Association.

Student registration is only $80. The non-student registration fee is $310.

High-quality workshops (described below) are also available, two of which are free of charge (thanks to support from the Agency for Healthcare Research and Quality and the National Center for Health Statistics), with the others available for $60 (non-students) or $30 (student rate).

The conference Web site is
<http://www.amstat.org/meetings/ichpr/2005/index.cfm>.

ICHPR INVITED SESSIONS
The following invited paper sessions are scheduled:
+Predicting High-Cost Users of Medical Care and the Persistence of High Expenditures Over Time
Speakers: John Fleishman (AHRQ), Steve Cohen (AHRQ), Joel Cohen (AHRQ)

+Statistical Issues in the Hospital CAHPS (HCAHPS) Survey
Speakers: Paul Cleary (Harvard Medical School), Marc Elliott (RAND), James O'Malley (Harvard Medical School)
Discussant: Ron Hays (RAND)

+Imputation in High-dimensional Complex Surveys
Speakers: Recai Yucel (University of Massachusetts-Amherst), Fabrizia Mealli (University of Florence), Nathaniel Schenker (NCHS)

+Assessing Pharmaceutical Safety and Efficacy in the Wake of COX-2 and HRT
Speakers: Muhammad Mamdani (Institute for Clinical Evaluative Sciences, Toronto), Alan Breier (Chief Medical Officer, Eli Lilly and Company), Robert O'Neill (FDA)
Discussant: Frank E. Harrell, Jr. (Vanderbilt University)

+Methods in Longitudinal Data Analysis
Speakers: Liming Cai (NCHS), John Kautter (RTI), Geert Molenberghs (Limburgs Universitair Centrum, Belgium)

+Methods of Risk Adjustment for Skewed Outcome Data
Speakers: Alberto Holly (University of Lausanne), Kenneth Pietz (Houston VA Medical Center), Anirban Basu (University of Chicago)

+Population Needs Based Funding Models (International Health Policy Session)
Speakers: Peter C. Smith (University of York), Therese Stukel (ICES, Toronto), Peter Cramptom (University of Otago, New Zealand)
Discussants: Rob Reid (Univ. of British Columbia), Cameron Mustard (Univ. of Toronto)

+Advanced Methods for Estimating Health Disparities
Speakers: Kathleen Cagney (Univ. of Chicago), Julianne Souchek (Houston VA Medical Center), Alan Zaslavsky (Harvard Medical School)

+Selection Bias in Observational Studies
Speakers: Joseph Terza (Univ. of Florida), Matthew Maciejewski (Univ. of Washington), Sharon-Lise Normand (Harvard Medical School)
Discussant: Douglas Staiger (Dartmouth College)

+Causal Inference with Longitudinal Data
Speakers: Babette Brumback (Univ. of Florida), Dylan Small (Univ. of Pennsylvania), Mary Beth Landrum (Harvard Medical School)

+Combining Estimates/Information Using Multiple Data Sources
Speakers: Michael Elliott (Univ. of Pennsylvania), Dawei Xie (Univ. of Pennsylvania), Nicholas Horton (Smith College)

ICHPR WORKSHOPS: Workshops 6&9 FREE, Others $60 (regular), $30 (student)

1. Statistical Graphics for Exploring Data, Presenting Information, and Understanding Statistical Models (Part I); Frank E. Harrell, Jr., Vanderbilt University; Friday, Oct. 28, 2005, 8:30 a. m. - 10:15 a.m.

2. Statistical Graphics for Exploring Data, Presenting Information, and Understanding Statistical Models (Part II); Frank E. Harrell, Jr., Vanderbilt University; Friday, Oct. 28, 2005, 10:30 a.m. - 12:15 p.m.

3. Strategies for Using Propensity Scores Well; Thomas E. Love, Case Western Reserve University; Friday, Oct. 28, 2005, 10:30 a.m. - 12:15 p.m.

4. Modern Meta-Analysis; Christopher H. Schmid, Tufts University School of Medicine; Friday, Oct. 28, 4:15 p.m. - 6:00 p.m.

5. Privacy, Confidentiality, and Data Security Training for Health Services Research; Alan M. Zaslavsky, Harvard University; Friday, October 28, 4:15 p.m. - 6:00 p.m.

6. Research Opportunities Using AHRQ Databases; Karen Beauregard, Agency for Healthcare Research and Quality; Saturday, Oct. 29, 8:30 a.m.-10:15 a.m. (NO WORKSHOP FEE)

7. Advances in Latent Variable Modeling (Part I); Bengt Muthen, University of California, Los Angeles; Saturday, Oct. 29, 8:30 a.m.-10:15 a.m.

8. Advances in Latent Variable Modeling (Part II); Bengt Muthen, University of California, Los Angeles; Saturday, Oct. 29, 10:30 a.m.-12:15 p.m.

9. Research Opportunities Using Data from the National Center for Health Statistics; Jim Lubitz and Robert Weinzimer, National Center for Health Statistics; Saturday, October 29, 10:30 a.m. - 12:15 p.m. (NO WORKSHOP FEE)

10. Issues when Using Hierarchical Models To Estimate Provider Performance; Michael Shwartz and Arlene Ash, Boston University; Saturday, October 29, 2:00 p.m. - 3:45 p.m.

11. Risk Adjustment and Predictive Modeling; Randall P. Ellis, Boston University; Saturday, Oct. 29, 4:00 p.m. - 5:45 p.m.

12. Bayesian Hierarchical Modeling with Applications to Provider Profiling (Part I); David Draper, University of California, Santa Cruz; Sunday, Oct. 30, 8:30 a.m. - 10:15 a.m.

13. Bayesian Hierarchical Modeling with Applications to Provider Profiling (Part II); David Draper, University of California, Santa Cruz; Sunday, Oct.30, 10:30 a.m. - 12:15 p.m.

134th Annual Meeting Abstracts Due February 2006

The 134th Annual Meeting of the American Public Health Association will be held in Boston, Nov. 4-8, 2006, with the theme: Public Health and Human Rights. Members are reminded that abstracts will be due in February 2006.

Anyone interested in organizing an Invited Session or being a speaker at an Invited Session should submit a proposal to the program chair.

The proposal should include
1. the organizer's full mailing address, e-mail address, telephone and fax numbers;
2. session title;
3. a brief (1-2 pages) overview describing the purpose, relevance and importance of the proposed session;
4. topics and participants, including the proposed presiders, titles of each presentation, full names of all authors, with full names of presenters underlined (as in the online abstract); and
5. a time schedule within the session (begin the session at 0:00 hrs and end it at 1:30).
A presentation at a Special Interest Session may be up to 30 minutes in length. We suggest that Special Interest Sessions allow time for a discussion. All Special Interest Session proposals and abstracts will be peer reviewed.

Submit the proposal to the 2006 Program Chair:
G. David Williamson
Director
Agency Toxic Substances & Disease Registry (ATSDR)
Office Assistant Administrator, Division of Health Studies
Centers for Disease Control and Prevention
2400 Century Pkwy
Atlanta, GA 30345
dxw2@cdc.gov
(404)498-0105
Fax: (404)498-0077