Bailar John C. - Medical Uses of Statistics

Copyright 2009 by Massachusetts Medical Society. All rights reserved.

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.

Published simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400, fax 978-750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, Massachusetts Medical Society, 860 Winter Street, Waltham, MA 02451.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representation or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor the author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at 877-762-2974, outside the United States at 317-572-3993 or fax 317-572-4002.

Wiley also published its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging in Publication Data:

Medical uses of statistics/edited by John C. Bailar III, David C. Hoaglin. 3rd ed.

p.; cm.

Includes articles originally published in the New England journal of medicine.

Includes bibliographical references and index.

ISBN 978-0-470-43952-4 (cloth) ISBN 978-0-470-43953-1 (pbk.)

1. Medical statistics. 2. Clinical medicine Research Statistical methods. I. Bailar III, John C. (John Christian), 1932- II. Hoaglin, David C. (David Caster), 1944- III. New England Journal of Medicine.

[DNLM: 1. Statistics as Topic Collected Works. 2. Research methods Collected Works. WA 950 M489 2009]

RA409.M43 2009

610.72dc22

2009017256

To

Frederick Mosteller (19162006)

superb teacher

supportive friend

and wise collaborator

Contributors

Shilpi Agarwal, M.B.B.S.

Department of Epidemiology; Harvard School of Public Health

Paul S. Albert, Ph.D.

Biometric Research Branch, National Cancer Institute

John C. Bailar III, M.D., Ph.D.

Professor Emeritus, University of Chicago; Scholar in Residence, National Academies

A. John Bailer, Ph.D.

Department of Mathematics & Statistics, Miami University

Graham A. Colditz, M.D., Dr.P.H.

Department of Surgery, Washington University School of Medicine

Fernando Delgado, M.S.

Colombia, South America

Christi Donnelly, D.Sc.

Department of Biostatistics, School of Public Health, Harvard University

Jeffrey M. Drazen, M.D.

Editor-in-Chief, New England Journal of Medicine

John D. Emerson, Ph.D.

Department of Mathematics, Middlebury College

Mark S. Goldberg, Ph.D.

Department of Medicine, McGill University

David C. Hoaglin, Ph.D.

Abt Bio-Pharma Solutions, Inc.

Hossein Hosseini, Ph.D.

Digital Equipment Corporation, Irvine, California

David J. Hunter, M.B.B.S.

Department of Medicine, Brigham & Womens Hospital; Harvard School of Public Health

Joseph A. Ingelfinger, M.D.

Bowdoin Street Health Center, Harvard Medical School

Thorsten Kurz, Ph.D.

Core Facility Genomics, University Hospital Freiburg, Germany

Stephen W. Lagakos, Ph.D.

Department of Biostatistics, School of Public Health, Harvard University

Philip W. Lavori, Ph.D.

Department of Psychiatry and Human Behavior, Brown University

Thomas A. Louis, Ph.D.

Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health

Nancy E. Mayo, Ph.D.

Division of Clinical Epidemiology, Department of Medicine, McGill University

Stephen Morrissey, Ph.D.

New England Journal of Medicine

Lincoln E. Moses, Ph.D. (19212006)

Department of Statistics, Stanford University

Frederick Mosteller, Ph.D. (19162006)

Department of Statistics, Harvard University

Dan L. Nicolae, Ph.D.

Department of Medicine and Department of Statistics, University of Chicago

Carole Ober, Ph.D.

Department of Human Genetics, University of Chicago

Margaret Perkins, M.A.

New England Journal of Medicine

Marcia Polansky, D.Sc.,

Department of Biometrics and Computing, Drexel University

Amita Rastogi, M.D., M.H.A.

Ingenix, Inc.

Paul J. Rathouz, Ph.D.

Department of Health Studies, University of Chicago

Michael A. Stoto, Ph.D.

School of Nursing & Health Studies, Georgetown University

Rui Wang, Ph.D.

Biostatistics Center, Massachusetts General Hospital; Department of Biostatistics, Harvard School of Public Health

James H. Ware, Ph.D.

Department of Biostatistics, School of Public Health, Harvard University

Preface

The practice of medicine combines science and art. The science part of medicine derives largely from inferences drawn from experiments, often performed with the invaluable assistance of patients who put themselves at risk to become research participants. These brave and altruistic people have all or part of their medical care driven by the requirements of research participation rather than by their specific clinical needs. Investigators measure various outcomes and assemble the results of their observations in research reports, which medical journals review and publish to help guide the communitys thinking about how best to approach the biology, prevention, diagnosis, and treatment of the condition under study.

It comes as no surprise that the clinical and laboratory observations involve many sources of variation, including measurement errors, intrinsic patient biological variability, and differences among patients in adherence to treatment protocols. These multiple sources of variation lead to uncertainty in assessments of outcome and in the clinical inferences drawn from them. Medical researchers apply statistical methods to these inherently noisy data and derive reasonably precise conclusions from them, taking into account not only the uncertainty but also other limitations of the data. Their experience with this process and its results also guides them in designing new studies. The conclusions drawn from these inferences drive clinical practice.