A S A P
American Statistical Association - Philadelphia Chapter
MEETING with ASA President
DATE: Wednesday, September 22, 1999LOCATION:
1:00 -- Introductory Remarks
Tad Archambault, Ph.D. (President, ASAP)
1:10 -- Joseph F. Heyse, Ph.D.
Measuring a Clinically-Meaningful Change in Quality of Life: Quantitative Perspective.
2:10 -- Jonas H. Ellenberg, Ph.D.
Communication of Statistical Concepts: Examples in Medical Collaboration.
All entrees include: Tossed Garden salad, rolls, butter, Chef's selection of vegetables and starch, Apple Crisp, choice of iced tea, brewed decaffeinated coffee, coffee or tea.
Please select one:
PAYMENT: Payment by cash or check at the meeting. Make check payable to ASAP.
RESERVATIONS: Please RSVP to Maria James with selection of lunch entrée by phone (610-397-7063) or fax (610-397-7105) or e-mail (firstname.lastname@example.org) by 4 pm, Tuesday, September 14, 1999. Cancellations cannot be accepted after that date, and "no shows" will be billed.
DIRECTIONS TO THE EMBASSY SUITES HOTEL (Airport)
Take Exit #10 (Philadelphia International Airport). Stay in your right-hand lane at all times. There will be another exit. This exit is for Route 291, Lester/Cargo City. Follow the ramp. It wraps around the Embassy Suites Hotel. The hotel is on your right.
Take Exit #10 (Philadelphia International Airport). Continue straight to your 1st traffic light. This is Bartram Avenue. Make a left onto Bartram Avenue. Continue straight for approximately one mile, the Embassy Suites Hotel is on your right.
SEPTA R1 Yellow Line to Airport
Cross over to the baggage claim area. Use the hotel courtesy phone for pickup. Then proceed outside to Zone 4.
Joseph F. Heyse, Ph.D.
Senior Director, Merck Research Laboratories
Clinical trials are an important source of information used by healthcare decision makers to establish health policy, to approve new drugs and vaccines for marketing, and to make judgments about prescribing treatments for specific patients. With these expanding information needs, patient perceptions of health-related quality of life (HRQOL) have received a higher priority in clinical evaluations of therapeutic and preventive interventions. In this paper, we will discuss the quantitative issues with evaluating clinically meaningful changes in HRQOL as they relate to the design, analysis and interpretation of clinical trials. The minimal clinically important difference (MCID) is generally defined as the difference in score on a HRQOL instrument that corresponds to the smallest change in health status that patients detect. Various measures of effect size have also been used as a measure of the clinical relevance of changes in HRQOL. We will discuss how the MCID and effect size can be estimated and how they are used in designing and interpreting trial results. It is also important to consider the multiplicity aspects of HRQOL in clinical studies. Multiplicity arises due to the essential multidimensionality of HRQOL instruments along with the usual design features of clinical trials (e.g., multiple hypotheses, timepoints, treatment comparisons, etc.).
We will discuss the quantitative aspects of measuring HRQOL using two trials. The first was a dose ranging study of a developmental asthma compound that included a four domain disease specific QOL instrument. The second study was a QOL evaluation that was part of ACTG320 to see whether adding the protease inhibitor indinavir to double therapy using zidovudine and lamivudine for HIV patients had a negative impact on QOL. We will compare and contrast the objectives and statistical methods used for these two very different studies.
* Joint work with George Carides and John Cook
Jonas H. Ellenberg, Ph.D.
President, American Statistical Association
Vice President and Senior Biostatistician, Westat, Rockville, MD
The statistician as a collaborator is often placed in the position of explaining or introducing well-understood statistical design concepts to audiences that may not be receptive. The lack of enthusiasm may be a function of a particular group's background in research design, and/or the statistician's inability to present the approach in a language familiar to the intended audience. It is posited that the use of examples relevant to the audience being addressed is an effective communication tool for getting collaborative groups to "buy in" to fundamental design principles. Several examples are given that demonstrate the impact of sample selection bias on the success and failure of previous medical studies.