Also in this issue, Renee Stolove, Director of Clinical Education in the New York Medical College Program in Physical Therapy, gives us her thought- provoking "outsider's" perspective on instruction in basic biostatistics and epidemiology.
This issue features another installment of a regular feature of this Newsletter - reviews of biostatistics textbooks by instructors who are currently using them. This issue's review is by Yvonne Michel of the Medical University of South Carolina College of Nursing and the book is the 2nd edition of David Moore's The Basic Practice of Statistics. If you have suggestions for future books for review, please contact Joe Shih, our Book Review Editor. His e-mail address shihwj@umdnj.edu
Also contained in this issue are messages from the 2000 and 2001 Program Chairs for the Section, Bob Oster and Lynn Ackerson. Bob gives us a wrap-up for the 2000 JSM in Indianapolis, and Lynn gives us a preview of the 2001 JSM in Atlanta. Pay particular attention to Lynn's call for your participation in the 2001 JSM on p. 5 of the Newsletter.
Finally, please note the announcements at the end of the Newsletter. One is for an undergraduate research program in biostatistics to be held at The Ohio State University during summer 2001. This is a great opportunity for your undergraduate students interested in a career in biostatistics to gain some practical experience. The other announcement is a request from the Director of the Division of Transplantation in the U.S. Dept. of Health & Human Services for nominations for the Secretary's Advisory Committee on Organ Transplantation. If you know of someone that would be interested in serving on this Committee, contact information is provided on p. 7 of the Newsletter.
Stephen Looney
A small group may vary in size from 7 or 8 students to as many as 35 to 40 students. It has been suggested that the size of a group should not exceed 10 if there is to be an expectation that all students will be active participants in discussion. For many years the size of the small groups in the Epidemiology and Biostatistics course at Mount Sinai School of Medicine were 19 or 20 students. Beginning in 1992, the size was decreased to 9 or 10, and the change in group dynamics was as one might expect. Instead of being a smaller version of an instructor-centered lecture session, these sessions became a place where students were freer to participate both by asking questions and by contributing their ideas.
The American Association of Medical Colleges (AAMC) instituted an online data management system, the Curriculum Management and Information Tool (CurrMIT), in the summer of 1998. The database contains descriptions of the courses taught, the total number of hours per course, the educational methods used, and the number of hours in each of the various formats for each course.
I did a search in CurrMIT on the word "biostatistics". There were 34 medical schools with data in the database on their biostatistics courses. The median number of total hours in these courses is 28, with a range from 12 to 49 hours; and the median number of hours devoted to lecture is 18, with a range from 2 to 43. For hours devoted to small group sessions, the median is 5 hours, and the range is from 0 to 31.
The small group, particularly when it contains no more than 10 students, is an excellent place either to introduce new concepts or to reinforce concepts introduced in lecture. The format of the teaching exercises in small groups can take many forms. Our experience has shown that varying the format works well. In 1983, Howard Gardner, a professor of education at Harvard, proposed the idea of multiple intelligences. Basically this theory says that individuals learn through a variety of approaches. In our experience, we have found that some students learn better from one format than from another, and some tutors feel more comfortable with one format than with another. The formats that are used in our course in epidemiology and biostatistics are the following: traditional problem sets, problem-based learning (PBL) cases, and evidence-based medicine (EBM) cases.
The traditional problem set format is one based on either statistical problems to solve using a standard test, such as chi-square, or a paper from the medical literature with an accompanying set of questions about the statistical methodology used in the paper. This format is the most structured of all the formats used, and it is the one with which our students seem to be most comfortable. All of the other courses they have had at the time they take the biostatistics course have used this format for any small group sessions that may have been included in the course. Because of the structure, the students feel it is an efficient use of their time and that they are maximizing the content to be learned.
The PBL format is a format used extensively at the Harvard Medical School. In this format, a "case" consists of a scenario in several parts. There are no structured questions provided to the students. They read the narrative of the case in the small group session; and, as a group, they decide what the issues are in the material they have just read. They formulate their own learning agenda, and between sessions do outside reading and research to answer the questions they themselves raised. In our experience, the PBL case encountered in the epidemiology/biostatistics course is the first PBL format they have ever seen. As a result, there is a general level of discomfort with the unstructured nature of the case, and the students feel they are wasting valuable time in trying to figure out what to learn and that it would be much more efficient if only the faculty would tell them what are the important things to be learned.
Under the EBM format in our course, each small group is given a particular question and each student in the group is asked to find a single paper from the recent literature relating to the question. The students critically read their paper before class, and in the small group session there is a general discussion of the assigned question with students contributing to the discussion from what they have read. An example of a question used for one of these sessions is the following: Does residential exposure to indoor radon increase the risk of developing lung cancer? This format has been used in our course as the last seminar session when the students have learned how to critically read papers from the literature. They seem to enjoy this session and they find it fun to be able to read papers and identify which are the stronger and weaker ones and to synthesize the results from several papers on a single topic.
At the end of the course the students were asked to evaluate each of the three small group formats on a seven-point scale, ranging from 1="poor, learned little" to 7="excellent, learned a great deal". The results were that favorable evaluations were given to all three formats, with more than half of responders giving ratings above 4. Finally, the students were asked for a preference rating of Traditional format vs. PBL format. The results were that there was a moderate preference of the Traditional.
In conclusion, the small group experience is a way to reinforce material presented in lecture and to be able to respond to students’ questions in a way that is not possible in a large lecture format. It also provides for an active learning experience on the part of the students. The CurrMIT data show that many courses in medical schools have small groups as a component, but many courses are predominantly still lecture. By using different formats for presenting material in the small group session, variety is introduced and most students will find a style that works for them.
If you participate in medical education, particularly as a course director, I encourage you to check the CurrMIT data for your school. This can be done by going to the website for the Association of American Medical Colleges http://www.aamc.org. If those of us involved in medical education will contribute the relevant data to this database, it will be a tremendous resource for the exchange of ideas.
Presenting a poster at Indianapolis took a great deal of nerve. Not only did I feel completely intimidated by the audience, but I had no "hard data" to present. Where would I find the unmitigated gall to present anecdotal results to (gasp) real statisticians.
Happily, my fears were completely unfounded. I want to thank everyone for the warm welcome at JSM 2000. Furthermore, it seemed that the results of efforts to teach statistics to future healthcare workers were about the same everywhere. We can restructure course curriculum, or in Sara Vesely's case, upgrade the level of the entire program. At best, students will merely go through the motions and regurgitate information to enable them to pass your exams. When it comes to a true understanding of basic statistics, our students just don't get it and even worse, they just don't seem to care. This leads one to a realization of both the true importance and major weakness in teaching statistics to future healthcare professionals.
Healthcare currently exists in a dynamic environment. It is changing from an art to a true science, with an increasing emphasis on evidence-based practice. More than ever, healthcare workers need the ability to critically analyze the available evidence to optimize their clinical decisions. As educators, it is our responsibility to prepare these future practitioners to do this. One would think that now, more than ever, these future practitioners need explicit and rigorous training in Statistics and Research Design.
Healthcare educators need to clarify just what the desired outcomes of entry-level professional education are. With regard to Statistics, just what do we expect our students to be able to do. Reading and interpreting the literature would appear to be a simple and obvious answer. At New York Medical College's PT Program, we certainly thought that we were on the right track. Students were required to take one term of undergraduate statistics as a prerequisite course. A Research Design course was given the term before their Problem-Based learning courses in Clinical Science. This did not work out as well as intended because:
Healthcare educators must also share the responsibility for teaching this subject matter to their future clinical colleagues. We cannot expect statisticians to fully appreciate the realities of the clinic. Collaborative efforts between departments are needed; particularly to develop relevant curricular activities. This was exemplified at the poster session in Indianapolis by Sondra Perdue and Ruth Rea, who do co-teach. (A sense of humor doesn't hurt, either!)
Finally, we have to stop relegating this subject matter to only the courses that are supposed to teach it. Isn't this the ultimate message that these courses have nothing to do with clinical practice? Statistics and research issues can be integrated into the clinical curriculum. I have implemented this on a small scale this year. For example, when our students started learning clinical assessment techniques, they received a lecture on reliability and validity of measurement. During one of their clinical labs, we will have them gather data to assess the inter and intra-rater reliability of a particular measurement. Dare I request that the students analyze the data? If none of them have asthma attacks when this is suggested, why not? (Note: some of our students have developed nosebleeds when trigonometric functions were mentioned during their Biomechanics course. As a healthcare worker, I must uphold the principle of non-malfeasance.)
I look forward to letting everyone know
how all of this turns out this summer in Atlanta.
Health Sciences students are demonstrably competent learners. Give them a useful resource, and they will devour it. Give them a statistical textbook that speaks about generating random numbers and probability density functions, and it will be dutifully read, but since they don't have probability or calculus backgrounds, it will not be integrated into their knowledge base. Relevant examples are useful. Formulas are not.
I have collected over 50 statistical texts designed specifically for health sciences and/or non-statisticians. Students regularly complained about the non-readability of the texts I selected for class. In despair, I distributed my cache of textbooks to the students for their review and evaluation as a credit-generating exercise. Students were also encouraged to identify other statistics books from any source. Strengths and weaknesses were cited for the texts and recommendations were made about using the book in the next class.
The book that was chosen was David Moore's The Basic Practice of Statistics, 1st Edition. Strengths cited by students were that it was reasonably readable, had a sufficient number of relevant examples, had a strong data analysis component, and was written such that the chapter on probability was optional. There was keen competition among the top texts, but BPS was the overall favorite although students viewed the exercise as identifying the least of all statistical devils.
As it happened during the following summer, the 2nd Edition of BPS was published with companion CD-ROM, Test Bank, Student Study Guide, and software manuals (Minitab, Excel, SPSS, and SAS). Respecting my prior students' preferences, I used this 2nd edition of BPS with the New Excel Manual the following semester.
The BPS was well received by the next class. They still had to work hard to comprehend the material, but they rated the textbook as good. This might sound like condemnation through faint praise, but the fact that these students rated a stat book positively is, in my experience, astonishing. Their review was not without complaint, though. Students found it difficult to access the chapters/sections on the CD, some data for the exercises were not available, the Q & A's on the CD were too easy/too few, and many figures in the book were without interpretable axes and frames of reference.
Students tolerated the Internet Testing Service using the Text Bank rather well and, after the course was complete, recommended its continued use. Students rated that the New Excel Manual as limited in its usefulness and I dropped it from the course’s required book list.
I still collect statistical textbooks. I
continue to have students review the books. BPS
has not lost a competition yet.
The TSHS invited session is already planned and it sounds very exciting. We will be having a debate: "Resolved: Required Biostatistics courses should be dropped from medical school curricula." Those arguing in support of the resolve will be Naomi Fineberg from Indiana University and Ted Colton from Boston University. Those arguing against the resolve will be Beth Dawson from Southern Illinois University and Peter Imrey from the University of Illinois. Dan Freeman of the University of Texas Medical Branch at Galveston is the organizer of this debate and will act as the moderator.
I would like you to consider getting involved in JSM 2001 by making a contributed presentation of some form. As has been done previously, TSHS will be giving awards for Best Invited Paper, Best Contributed Paper, and Best Contributed Poster. Abstracts for contributed presentations are due February 1, 2001. Information and forms are available in the November issue of Amstat News, and an online version should be available by Dec. 1 at http://www.amstat.org/meetings/jsm2001/abstracts. I would highly recommend that you use the online method if at all possible. Note that virtually all submitted abstracts are accepted for the meetings.
As you consider the above, recall that the mission statement of TSHS states that we are "...devoted to excellence in teaching statistical methods and basic epidemiology and in statistical consulting within the health sciences". If this mission covers most of your professional work, please focus your presentation towards this mission, and then designate your JSM abstract as being for the TSHS section.
Authors of contributed presentations must choose from one of the following four formats: Topic Contributed Papers (formerly Special Contributed Papers), Topic Contributed Panel (formerly Special Contributed Panel), Regular Contributed Papers, and Regular Contributed Posters. Topic Contributed Paper sessions consist of five 20- minute presentations arranged in advance by an organizer. Topic Contributed Panels consist of three to five panelists providing commentary on a topic arranged in advance by an organizer. Regular Contributed Paper sessions consist of five to seven 15-minute presentations that are clustered around a general theme. Regular Contributed Poster sessions run for two hours and permit extended face-to-face discussions with attendees who are particularly interested in the topic. If you are interested in organizing a Topic Contributed Session or a Topic Contributed Panel, please contact me as soon as possible (LMA@dor.kaiser.org).
Furthermore, all invited and contributed papers and posters in TSHS sessions are eligible for publication in the 2001 ASA Proceedings volume that we share with the Section on Statistical Education and the Section on Statistical Consulting. Contributing to the proceedings is voluntary. If you decide to contribute to the proceedings, note that there is no peer or editorial review. In addition, you can still publish your paper in a regular peer- reviewed journal.
Your participation in TSHS sessions at JSM
2001 will be greatly appreciated. Feel free to
contact me if you have any questions or concerns. I
look forward to seeing you in Atlanta!
My best estimates of attendance at each session follow. These estimates include people who may have come for one or two talks and then left.
The winners of the Best Invited Paper Award are Stephen Looney and Prasaad Steiner, both of the University of Louisville School of Medicine. Their presentation was entitled "How Should We Teach Basic Biostatistics and Epidemiology to Medical Students?".
The winners of the Best Topic Contributed Paper Award are Rebecca Moore, of the Oklahoma Department of Mental Health and Substance Abuse Services, and Barbara Neas, of the Oklahoma University Department of Biostatistics and Epidemiology. Their presentation was entitled "Disseminating Information – Tales to Tell About Treatment Outcomes".
The winners of the Best Regular Contributed Paper Award are Michael Mosier, of the University of Kansas Medical Center, and James Whitmore, of the Immunex Corporation. Their presentation was entitled “An In-Class Exercise for Teaching Clinical Trial Design Issues to Health Science Students and Professionals”.
The winners of the Best Regular Contributed Poster Award are Sondra Perdue and Ruth Rea, both of the University of Washington - Tacoma. Their presentation was entitled “Dead Numbers Society: Helping Students Bring Dead Numbers to Life”.
Congratulations to all of the above award winners!
Some of the JSM 2000 TSHS papers may be archived on the TSHS Web site http://www.bio.ri.ccf.org/ASA_TSHS/, which can also be accessed via a link from the ASA’s webpage. In addition, some of these papers may appear in the JSM 2000 Proceedings volume that we share with the Section on Statistical Education and the Section on Statistical Consulting.
Thanks again to all the many people who helped make the TSHS offerings such a big success in Indianapolis. As you begin to look toward JSM 2001, please give your ideas and suggestions to Lynn Ackerson, our Program Chair for 2001. I hope that we will have a full program for JSM 2001, with participation by many section members.
In conclusion, I have enjoyed serving TSHS
during the past two years (this year as Program Chair
and last year as Program Chair-Elect). It has truly
been a wonderful experience. I have very much
enjoyed working with each and every one of you. ?
Each participant will be assigned to a project requiring the application of statistical methods to real problems. During the first 5 weeks of the program a participant will take intensive courses in applied statistics appropriate for her/his previous exposure to the field. In these courses, they will learn to reason statistically, use statistical software, manage data, and perform standard statistical analyses. Throughout the entire 8 weeks each participant will meet with faculty mentors concerning the subject-matter issues related to their specific project, as well as with statistics faculty, staff, and graduate research associates to discuss alternative ways of analyzing the data associated with their project. Each participant will prepare a written report documenting the relevant scientific questions and the associated statistical solutions for their project and make a verbal presentation of their report to the other participants. A poster session will be organized at which the program participants can also share their summer experiences with the scientific community at The Ohio State University.
Additional information about this undergraduate research program in biostatistics can be found at our Department of Statistics website http://www.stat.ohio-state.edu/. If you wish to inquire further about the summer program or have questions, please send an e-mail to reu@stat.ohio-state.edu. The deadline for receipt of completed applications is February 1, 2001. ?
Douglas A. Wolfe
Department of Statistics
The Ohio State University
This committee is being established in accordance with the Organ Procurement and Transplantation Network Amended Final Rule, which authorizes the Secretary to establish the Committee, which shall include members that are knowledgeable in such fields as health care public policy, transplantation medicine and surgery, non-physician transplant professions, biostatistics, immunology, health economics, epidemiology, and bioethics, as well as representatives of transplant candidates, transplant recipients, organ donors, and family members.
If you wish to submit a nomination, contact Dr. Miguel Kamat at (301) 443-8712. The mailing address is
Department of Health & Human Services
Health Resources and Services Administration
Office of Special Programs
Division of Transplantation
Rockville, MD 20857
Be sure to visit the Teaching Statistics in the Health Sciences website http://www.bio.ri.ccf.org/ASA_TSHS/. A new addition is a link to files related to a recent article by our Section President:
O'Brien, RG. Applying for a Job: Your Curriculum Vitae and Cover Letter, Amstat News, September 2000.
These files include a downloadable copy of Ralph’s article and MS Word and FrameMaker templates for preparing a CV according to his recommendations. Even the most experienced professional among us could stand to improve his or her CV and Ralph’s article will help us do just that. Check it out!
Stephen Looney
| ||||||||||||||||||||||||||||||||||||||||||||
Best Invited Paper: Stephen Looney and Prasaad Steiner
Best Topic Contributed Paper: Rebecca Moore and Barbara Neas
Best Regular Contributed Paper: Michael Mosier and James Whitmore
Best Regular Contributed Poster: Sondra Perdue and Ruth Rea
American Statistical Association
TSHS Newsletter
Stephen W. Looney, Ph.D., Editor
1429 Duke Street
Alexandria, VA 22314-3415
USA