Skip Navigation
AHRQ--Agency for Healthcare Research and Quality: Advancing Excellence in Health Care
  • Home
  • Methods for Developing and Analyzing Clinically Rich Data for Patient-Centered Outcomes Research

Third Symposium on Comparative Effectiveness Research Methods

Methods for Developing and Analyzing Clinically Rich Data for Patient-Centered Outcomes Research

Call for Abstracts for AHRQ Symposium
June 6 and 7, 2011

John M. Eisenberg Building
Agency for Healthcare Research and Quality Conference Center
Rockville, Maryland

Save this page in  Save this page in  Save this page in  Save this page on your Google Home Page  Save this page in Windows Live
Save this page in Yahoo  Save this page in  Stumble this page.  Save this page in  Digg this page. 

Print Print

The Agency for Healthcare Research and Quality, through its Effective Health Care Program, is sponsoring a third invitational symposium on research methods for comparative effectiveness (CER) and patient-centered outcomes research (PCOR). This 2-day symposium and workshop will be held on June 6 and 7, 2011 at the AHRQ Conference Center in Rockville, Maryland. The symposium is a follow-up to the 2006 and 2009 AHRQ conferences on Methods in Comparative Effectiveness and Safety Research; papers presented at past conferences were published in the journal Medical Care (copies may be downloaded from or These AHRQ methods symposia and the associated journal supplements have become important resources for investigators to use in designing and conducting comparative effectiveness studies.

The 2011 conference will focus on original research and the methodological insight or advances that arise from the conduct of the research. This includes methods development, enhancement, or application that would be appropriate for publication as a supplement in a peer- reviewed journal. Preference will be given to work that fits the theme of the symposium and 1) represents an innovative approach to CER/PCOR and the challenges of practice based research; 2) is directly applicable to comparative effectiveness (such as one designed to address a specific policy question or prioritized evidence gap); and 3) can be presented as a prelude to publication in a journal supplement that will be completed by early summer 2012. Authors of abstracts that are selected for the symposium will be expected to submit a complete manuscript for peer-review and publication in a special journal supplement of the symposium’s proceedings.

This announcement is an invitation for your participation in this third symposium, specifically through submission of a brief abstract on relevant research methods by March 1, 2011, to

More detailed information about the conference sessions and the abstract form are available in the next section.

Symposium Overview

Selecting the most appropriate data sources, research design, and analytic strategy must follow from a comparative effectiveness research question that is constructed to inform specific health care decisions. The 2011 Methods in Comparative Effectiveness symposium will examine new and emerging methods for comparative effectiveness and patient-centered outcomes research that either use clinically rich data sources or enrich existing data sources with supplemental data. Decisionmakers need valid information on the comparative effectiveness of medical products, interventions, diagnostics and delivery systems in routine care. Administrative data sources capture healthcare encounters longitudinally and serve as a comprehensive data backbone reflecting routine care, however, they often lack clinical detail or patient-centered outcomes, which may limit their utility for research intended for decisionmakers. We aim to explore methodologic options for enhancing secondary data sources or for prospectively designed registry studies and randomized trials that reflect routine care as much as possible to maximize utility for generating new scientific evidence that can be used for informing health care decisions.

The first half of the symposium will focus on addressing these issues within the context of observational studies based on secondary data enriched with clinical data while the second half of the symposium will focus on prospective randomized and observational approaches. It is anticipated that the abstracts chosen for presentation within each section will either describe a novel approach to addressing the overall meeting theme or focus on the application of an existing method to a particular problem using empirical data or simulation studies. Applications may include a wide range of CER questions, including the evaluation of delivery systems, medical products, procedures or interventions and diagnostics.

The two main emphases include examination of ways to enhance the study of treatments in comparative and clinical effectiveness studies by 1) enriching administrative data with clinically detailed patient information derived from electronic medical records, registries, in-patient databases, or surveys and by 2) innovative randomized study designs that mimic routine care.

Cutting across these are four thematic areas:

  1. study design
  2. data collection, linkage, and measurement
  3. statistics and analytic methods, and
  4. policy issues and applications

Applications may come from a wide range of relevant topics such as:

  • prescription drugs
  • medical devices
  • invasive and non-invasive procedures
  • complex clinical strategies
  • healthcare delivery systems and improvement programs, and
  • diagnostics

The listing below illustrates the range of topics AHRQ aims to cover, although it is neither comprehensive nor final. We are soliciting abstracts for approximately 15 plenary presentations; sessions will feature considerable time for discussion. Papers will be selected via a blinded, independent peer review process by members of the Scientific Planning Committee.

The organizers of the conference will coordinate the publication of a journal supplement (journal TBD) that will include manuscripts arising from the presentations at the conference. Conference presentation will greatly increase the probability of publication, but will not guarantee publication. To be considered for inclusion in this supplement, manuscripts arising from the conference must be submitted and revised according to publication timelines and subject to editorial and peer review.


Contact: Sebastian Schneeweiss, MD, ScD (
John Seeger, PharmD, DrPH (
Elizabeth Robinson, BA (, and/or
Scott R. Smith, PhD (

Scientific Planning Committee:

Scott R. Smith, PhD, MSPH (co-Chair, AHRQ)
Sebastian Schneeweiss, MD, ScD (co-Chair, Brigham & Women’s Hospital)
Alan Brookhart, PhD (University of North Carolina, Chapel Hill, Epidemiology)
Robert Glynn, PhD, ScD (Harvard School of Public Health, Biostatistics)
Eric Johnson, PhD, MPH (Kaiser NorthWest, Epidemiology)
Parivash Nourjah, PhD (AHRQ)
Jason Roy, PhD (University of Pennsylvania, Biostatistics)
John Seeger, PharmD, DrPH (Brigham & Women’s Hospital, Pharmepi)
Ravi Varadhan, PhD (Johns Hopkins University, Biostatistics)

Topics to be Considered

1. Improved identification of exposure, outcomes, and patient subgroups

While secondary healthcare data are often useful for studying drug exposures, other data sources may be needed to identify implantable devices (e.g., cardiac, orthopedic) or other forms of treatment (e.g., psychotherapy, type of surgery). The assessment of treatment effect heterogeneity relies on the accurate identification of clinically relevant patient subgroups by markers often not available in administrative databases (e.g., reduced cardiac ejection fraction when studying heart failure treatment). Clinically detailed data sources like electronic medical records can be mined to identify specific clinical diagnoses with high accuracy (e.g., hypertension, community acquired pneumonia).

This session will showcase practical experiences with linking intervention-specific detailed information to administrative data and demonstrate the utility of such combined data on effect estimation and subgroup analyses. It will provide the current state of outcome identification of specific clinical diagnoses in routine care data repositories, including supervised learning methods or natural language processing.

We specifically invite abstracts that describe empirical studies that describe analyses where exposure, outcomes, or patient subgroups were identified in information-rich clinical databases, e.g., electronic medical records or specific registries etc. with or without linkage to administrative data.

2. Improved confounder adjustment using supplemental data collection

Analyses conducted within existing healthcare databases are often challenged by a lack of detailed risk factor information that may represent a source of residual confounding. Supplementing these databases with detailed clinical information is feasible in some settings, allowing studies to address this limitation. However, new challenges may be introduced by such data, including appropriate use of the detailed confounder information that may be available for only a sample of the main study and appropriate methods for dealing with missing clinical detail, such as tests that may not have been ordered or reported.

In this session, we will learn about the extent of these new challenges and first-hand experiences with new techniques and analytic approaches that may eventually result in better confounding control in CER studies linking existing health care databases with supplemental secondary or primary data collections.

We specifically invite abstracts that describe empirical studies that have encountered and dealt with such potential confounding variables, and preference will be given to abstracts that feature novel approaches.

3. Randomized designs that mimic routine care

Decisionmakers request information on causal relationships between medicines and outcomes that are of highest validity and generalizable to routine care. Can randomized trials ever mimic routine care or are findings from RCTs never generalizable to routine care? We try to identify study questions and complementing randomized study designs that mimic routine care to an extent that they provide answers on the effectiveness medical products, interventions, and health care delivery systems that are generalizable to real world healthcare.

In this session we will learn about innovative trial designs and lessons from their implementation and empirical evidence on to what extent they represent diverse patient populations and real world adherence patterns.

As an example, the topic of large streamlined trials or cluster randomized trials could be presented. These studies address concerns about unmeasured baseline confounders and also address concerns about differences between clinical trial populations and patients who might receive a therapy as part of routine care.

We look forward to abstracts with empirical data from examples of innovative approaches or comparative applications.

4. Non-randomized prospective studies

Prospective assessment of outcomes may be necessary as secondary data often do not accurately assess selected exposures and more often not the right outcomes measures to answer the decisionmakers’ questions. Prospectively assessing confounding factor may further improve the validity of non-randomized studies.

In this session we hope to learn about novel approaches to efficient and targeted prospective studies in real world settings. Hybrid designs of administrative data, registries and original data collection have emerged as promising approaches.

In this session we will explore novel approaches to efficient and targeted prospective studies in real-world settings. Hybrid designs involving administrative data, registries and original data collection have emerged as promising approaches.

We look forward to abstracts with empirical data from examples of innovative approaches or comparative applications.

Instructions for Submission of Papers/Presentations for the Symposium

Please download the abstract submission form (29 kB) to provide the following information:

Page 1

  1. Title of presentation;
  2. Indicate research topic(s) or area(s) of interest (i.e., Topic category 1 thru 4 above or “other”);
  3. Identify the presenter and primary author of the subsequent manuscript (author and co-author names will not be included for blinded review of abstracts);
  4. Identify co-authors of the paper including their work affiliation or academic institution;
  5. Email address, phone number, and mailing address of primary author;
  6. Brief conflict of interest disclosure for primary author; and
  7. Brief conflict of interest disclosure for co- authors.

Page 2

  1. Title of presentation (this will allow matching with author names after blinded review);
  2. Describe your proposed paper in 350 to 500 words; and
  3. Indicate the stage of your research (e.g., already completed; will be completed by June 2011; can be completed by July 2011.

If you propose more than one paper, please send separate abstract submission forms for each paper.

Please send the completed abstract submission form to the Comparative Effectiveness Symposium email at Brigham and Women’s Hospital: DEcIDEMethods@partners.or g on or before March 1, 2011.

Please contact Sebastian Schneeweiss, MD, ScD (
John Seeger, PharmD, DrPH (
Elizabeth Robinson, BA (, and/or
Scott R. Smith, PhD (

On behalf of AHRQ's Effective Health Care Program, we welcome your participation and thank you for your interest.