Medical biostatistics

Tai ngay!!! Ban co the xoa dong chu nay!!! 16990153435761000000 Medical Biostatistics Fourth Edition Editor-in-Chief Shein-Chung Chow, Ph.D., Professor, Department of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham, North Carolina Series Editors Byron Jones, Biometrical Fellow, Statistical Methodology, Integrated Information Sciences, Novartis Pharma AG, Basel, Switzerland Jen-pei Liu, Professor, Division of Biometry, Department of Agronomy, National Taiwan University, Taipei, Taiwan Karl E Peace, Georgia Cancer Coalition, Distinguished Cancer Scholar, Senior Research Scientist and Professor of Biostatistics, Jiann-Ping Hsu College of Public Health, Georgia Southern University, Statesboro, Georgia Bruce W Turnbull, Professor, School of Operations Research and Industrial Engineering, Cornell University, Ithaca, New York Published Titles Adaptive Design Methods in Clinical Trials, Second Edition Shein-Chung Chow and Mark Chang Basic Statistics and Pharmaceutical Statistical Applications, Second Edition James E De Muth Adaptive Designs for Sequential Treatment Allocation Alessandro Baldi Antognini and Alessandra Giovagnoli Bayesian Adaptive Methods for Clinical Trials Scott M Berry, Bradley P Carlin, J Jack Lee, and Peter Muller Adaptive Design Theory and Implementation Using SAS and R, Second Edition Mark Chang Bayesian Analysis Made Simple: An Excel GUI for WinBUGS Phil Woodward Advanced Bayesian Methods for Medical Test Accuracy Lyle D Broemeling Bayesian Designs for Phase I–II Clinical Trials Ying Yuan, Hoang Q Nguyen, and Peter F Thall Analyzing Longitudinal Clinical Trial Data: A Practical Guide Craig Mallinckrodt and Ilya Lipkovich Bayesian Methods for Measures of Agreement Lyle D Broemeling Applied Biclustering Methods for Big and High-Dimensional Data Using R Adetayo Kasim, Ziv Shkedy, Sebastian Kaiser, Sepp Hochreiter, and Willem Talloen Bayesian Methods for Repeated Measures Lyle D Broemeling Applied Meta-Analysis with R Ding-Geng (Din) Chen and Karl E Peace Applied Surrogate Endpoint Evaluation Methods with SAS and R Ariel Alonso, Theophile Bigirumurame, Tomasz Burzykowski, Marc Buyse, Geert Molenberghs, Leacky Muchene, Nolen Joy Perualila, Ziv Shkedy, and Wim Van der Elst Bayesian Methods in Epidemiology Lyle D Broemeling Bayesian Methods in Health Economics Gianluca Baio Bayesian Missing Data Problems: EM, Data Augmentation and Noniterative Computation Ming T Tan, Guo-Liang Tian, and Kai Wang Ng Published Titles Bayesian Modeling in Bioinformatics Dipak K Dey, Samiran Ghosh, and Bani K Mallick Benefit-Risk Assessment in Pharmaceutical Research and Development Andreas Sashegyi, James Felli, and Rebecca Noel Benefit-Risk Assessment Methods in Medical Product Development: Bridging Qualitative and Quantitative Assessments Qi Jiang and Weili He Bioequivalence and Statistics in Clinical Pharmacology, Second Edition Scott Patterson and Byron Jones Biosimilar Clinical Development: Scientific Considerations and New Methodologies Kerry B Barker, Sandeep M Menon, Ralph B D’Agostino, Sr., Siyan Xu, and Bo Jin Biosimilars: Design and Analysis of Follow-on Biologics Shein-Chung Chow Biostatistics: A Computing Approach Stewart J Anderson Cancer Clinical Trials: Current and Controversial Issues in Design and Analysis Stephen L George, Xiaofei Wang, and Herbert Pang Causal Analysis in Biomedicine and Epidemiology: Based on Minimal Sufficient Causation Mikel Aickin Clinical and Statistical Considerations in Personalized Medicine Claudio Carini, Sandeep Menon, and Mark Chang Clinical Trial Data Analysis Using R Ding-Geng (Din) Chen and Karl E Peace Clinical Trial Data Analysis Using R and SAS, Second Edition Ding-Geng (Din) Chen, Karl E Peace, and Pinggao Zhang Clinical Trial Methodology Karl E Peace and Ding-Geng (Din) Chen Clinical Trial Optimization Using R Alex Dmitrienko and Erik Pulkstenis Cluster Randomised Trials: Second Edition Richard J Hayes and Lawrence H Moulton Computational Methods in Biomedical Research Ravindra Khattree and Dayanand N Naik Computational Pharmacokinetics Anders Källén Confidence Intervals for Proportions and Related Measures of Effect Size Robert G Newcombe Controversial Statistical Issues in Clinical Trials Shein-Chung Chow Data Analysis with Competing Risks and Intermediate States Ronald B Geskus Data and Safety Monitoring Committees in Clinical Trials, Second Edition Jay Herson Design and Analysis of Animal Studies in Pharmaceutical Development Shein-Chung Chow and Jen-pei Liu Design and Analysis of Bioavailability and Bioequivalence Studies, Third Edition Shein-Chung Chow and Jen-pei Liu Design and Analysis of Bridging Studies Jen-pei Liu, Shein-Chung Chow, and Chin-Fu Hsiao Design & Analysis of Clinical Trials for Economic Evaluation & Reimbursement: An Applied Approach Using SAS & STATA Iftekhar Khan Design and Analysis of Clinical Trials for Predictive Medicine Shigeyuki Matsui, Marc Buyse, and Richard Simon Design and Analysis of Clinical Trials with Time-to-Event Endpoints Karl E Peace Design and Analysis of Non-Inferiority Trials Mark D Rothmann, Brian L Wiens, and Ivan S F Chan Published Titles Difference Equations with Public Health Applications Lemuel A Moyé and Asha Seth Kapadia DNA Methylation Microarrays: Experimental Design and Statistical Analysis Sun-Chong Wang and Arturas Petronis DNA Microarrays and Related Genomics Techniques: Design, Analysis, and Interpretation of Experiments David B Allison, Grier P Page, T Mark Beasley, and Jode W Edwards Inference Principles for Biostatisticians Ian C Marschner Interval-Censored Time-to-Event Data: Methods and Applications Ding-Geng (Din) Chen, Jianguo Sun, and Karl E Peace Introductory Adaptive Trial Designs: A Practical Guide with R Mark Chang Joint Models for Longitudinal and Timeto-Event Data: With Applications in R Dimitris Rizopoulos Dose Finding by the Continual Reassessment Method Ying Kuen Cheung Measures of Interobserver Agreement and Reliability, Second Edition Mohamed M Shoukri Dynamical Biostatistical Models Daniel Commenges and Hélène Jacqmin-Gadda Medical Biostatistics, Fourth Edition A Indrayan Elementary Bayesian Biostatistics Lemuel A Moyé Meta-Analysis in Medicine and Health Policy Dalene Stangl and Donald A Berry Emerging Non-Clinical Biostatistics in Biopharmaceutical Development and Manufacturing Harry Yang Methods in Comparative Effectiveness Research Constantine Gatsonis and Sally C Morton Empirical Likelihood Method in Survival Analysis Mai Zhou Mixed Effects Models for the Population Approach: Models, Tasks, Methods and Tools Marc Lavielle Essentials of a Successful Biostatistical Collaboration Arul Earnest Modeling to Inform Infectious Disease Control Niels G Becker Exposure–Response Modeling: Methods and Practical Implementation Jixian Wang Modern Adaptive Randomized Clinical Trials: Statistical and Practical Aspects Oleksandr Sverdlov Frailty Models in Survival Analysis Andreas Wienke Monte Carlo Simulation for the Pharmaceutical Industry: Concepts, Algorithms, and Case Studies Mark Chang Fundamental Concepts for New Clinical Trialists Scott Evans and Naitee Ting Generalized Linear Models: A Bayesian Perspective Dipak K Dey, Sujit K Ghosh, and Bani K Mallick Handbook of Regression and Modeling: Applications for the Clinical and Pharmaceutical Industries Daryl S Paulson Multiregional Clinical Trials for Simultaneous Global New Drug Development Joshua Chen and Hui Quan Multiple Testing Problems in Pharmaceutical Statistics Alex Dmitrienko, Ajit C Tamhane, and Frank Bretz Published Titles Noninferiority Testing in Clinical Trials: Issues and Challenges Tie-Hua Ng Statistical Design and Analysis of Clinical Trials: Principles and Methods Weichung Joe Shih and Joseph Aisner Optimal Design for Nonlinear Response Models Valerii V Fedorov and Sergei L Leonov Statistical Design and Analysis of Stability Studies Shein-Chung Chow Patient-Reported Outcomes: Measurement, Implementation and Interpretation Joseph C Cappelleri, Kelly H Zou, Andrew G Bushmakin, Jose Ma J Alvir, Demissie Alemayehu, and Tara Symonds Statistical Evaluation of Diagnostic Performance: Topics in ROC Analysis Kelly H Zou, Aiyi Liu, Andriy Bandos, Lucila Ohno-Machado, and Howard Rockette Quantitative Evaluation of Safety in Drug Development: Design, Analysis and Reporting Qi Jiang and H Amy Xia Statistical Methods for Drug Safety Robert D Gibbons and Anup K Amatya Quantitative Methods for HIV/AIDS Research Cliburn Chan, Michael G Hudgens, and Shein-Chung Chow Quantitative Methods for Traditional Chinese Medicine Development Shein-Chung Chow Randomized Clinical Trials of Nonpharmacological Treatments Isabelle Boutron, Philippe Ravaud, and David Moher Randomized Phase II Cancer Clinical Trials Sin-Ho Jung Repeated Measures Design with Generalized Linear Mixed Models for Randomized Controlled Trials Toshiro Tango Sample Size Calculations for Clustered and Longitudinal Outcomes in Clinical Research Chul Ahn, Moonseong Heo, and Song Zhang Sample Size Calculations in Clinical Research, Third Edition Shein-Chung Chow, Jun Shao, Hansheng Wang, and Yuliya Lokhnygina Statistical Analysis of Human Growth and Development Yin Bun Cheung Statistical Methods for Clinical Trials Mark X Norleans Statistical Methods for Healthcare Performance Monitoring Alex Bottle and Paul Aylin Statistical Methods for Immunogenicity Assessment Harry Yang, Jianchun Zhang, Binbing Yu, and Wei Zhao Statistical Methods in Drug Combination Studies Wei Zhao and Harry Yang Statistical Testing Strategies in the Health Sciences Albert Vexler, Alan D Hutson, and Xiwei Chen Statistics in Drug Research: Methodologies and Recent Developments Shein-Chung Chow and Jun Shao Statistics in the Pharmaceutical Industry, Third Edition Ralph Buncher and Jia-Yeong Tsay Survival Analysis in Medicine and Genetics Jialiang Li and Shuangge Ma Theory of Drug Development Eric B Holmgren Translational Medicine: Strategies and Statistical Methods Dennis Cosmatos and Shein-Chung Chow http://taylorandfrancis.com Medical Biostatistics Fourth Edition Abhaya Indrayan Rajeev Kumar Malhotra CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2018 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S Government works Printed on acid-free paper International Standard Book Number-13: 978-1-4987-9953-9 (Hardback) This book contains information obtained from authentic and highly regarded sources Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint Except as permitted under U.S Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400 CCC is a not-for-profit organization that provides licenses and registration for a variety of users For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com Contents Summary Tables xxix Preface xxxv Frequently Used Notations xxxvii Medical Uncertainties .1 1.1 Uncertainties in Health and Disease 1.1.1 Uncertainties due to Intrinsic Variation 1.1.1.1 Biologic Variability 1.1.1.2 Genetic Variability 1.1.1.3 Variation in Behavior and Other Host Factors 1.1.1.4 Environmental Variability .3 1.1.1.5 Sampling Fluctuations 1.1.2 Natural Variation in Assessment 1.1.2.1 Observer Variability 1.1.2.2 Variability in Treatment Strategies 1.1.2.3 Instrument and Laboratory Variability 1.1.2.4 Imperfect Tools 1.1.2.5 Incomplete Information on the Patient 1.1.2.6 Poor Compliance with the Regimen .5 1.1.3 Knowledge Limitations 1.1.3.1 Epistemic Uncertainties 1.1.3.2 Chance Variability .6 1.1.3.3 Diagnostic, Therapeutic, and Prognostic Uncertainties 1.1.3.4 Predictive and Other Uncertainties 1.2 Uncertainties in Medical Research 1.2.1 Empiricism in Medical Research 1.2.1.1 Laboratory Experiments 1.2.1.2 Clinical Trials .7 1.2.1.3 Surgical Procedures 1.2.1.4 Epidemiological Research 1.2.2 Elements of Minimizing the Impact of Uncertainties on Research 1.2.2.1 Proper Design 1.2.2.2 Improved Medical Methods 1.2.2.3 Analysis and Synthesis .9 1.3 Uncertainties in Health Planning and Evaluation 1.3.1 Health Situation Analysis 1.3.1.1 Identification of the Specifics of the Problem 10 1.3.1.2 Magnitude of the Problem .10 1.3.1.3 Health Infrastructure .10 1.3.1.4 Feasibility of Remedial Steps 11 1.3.2 Evaluation of Health Programs 11 1.4 Management of Uncertainties: About This Book 11 1.4.1 Contents of the Book 12 1.4.1.1 Chapters 12 1.4.1.2 Limitations and Strengths .13 1.4.1.3 New in the Fourth Edition .14 1.4.1.4 Unique Contribution of This Book .14 ix 706 multiple linear regression, 426, 438, 443, 657 CI and tests in, 440 elements of, 438 multiple correlation in, 440 understanding, 439 multistage design, 301 multistage random sampling, 43, 46, 62, 585 multivariable analysis, 446 multivariate, 397, 529, 530, 561, 607 analysis, 412, 530, 537, 607 analysis of variance, 397, 536 categorical data, 337 comparisons, 151 conclusions, 604, 607 data, 146, 206, 545 entity, 205, 529 Gaussian distribution, 483, 540, 542 logistic, 471, 531 methods, 411, 529, 530, 532, 550, 574 multiple regression, 529, 531, 532 regression, 425, 538 results, 534 setup, 398, 412, 529, 530, 535, 538, 550, 586, 607 tests of significance, 535, 538, 628 mutually exclusive, 29, 39, 86, 99, 134, 183, 483 categories, 112, 137, 178, 182, 312, 316 events, 178 units, 39 N National Center for Health Statistics, 148, 149, 164 National Health Interview Survey, 23 national well-being, 246 natural clusters, 544, 545 number of, 545 natural experiment, 67 natural groups, 543 NCHS, see National Center for Health Statistics negative binomial distribution, 317, 635 negative predictivity, 18, 186–187, 197, 198, 265, 476 negative skewed, see left-skewed negative trial, 34, 101, 412 neonatal health, 160 neonatal mortality rate, 204, 229, 230 early, 204 late, 204 nested case–control studies, 59 nested design, see nested case–control studies net effect, 63, 68, 423, 424, 442 net reproduction rate, 221 New Zealand, 244 NNT, see number needed to treat n-of-1 trial, 94, 95, 100 nominal categories, 107, 317, 478, 483, 490, 497, 597 nominal scale, 107, 112, 129, 142, 208, 311, 332, 342 nomogram, 146, 299, 300 noncausal explanations, 35 noncompliance, 5, 20, 100, 577, 595 Index non-Gaussian conditions, 259, 271, 274, 401 non-Gaussian distribution, 266, 274, 401, 406, 570, 592 noninferiority margin, 326, 344 determining, 326 noninferiority trial, 92, 93, 267, 298, 325, 326, 595 nonlinear regression, 425, 427, 428, 438, 443, 662 nonlinear relationship, see nonlinear regression nonlinear score, 333 nonparametric method, 196, 197, 274, 276, 377, 391, 401, 414, 592, 594, 632 nonparametric test, 401, 405, 409 chi-square (distribution-free), see chi-square test Friedman, 407–409 Kolmogorov–Smirnov, 285, 401, 594 Kruskal–Wallis, 406 log-rank test, 517, 519, 521, 525 Mann–Whitney, 196, 404 sign, 401, 420 Wilcoxon test, 196, 401, 405, 519 Wilcoxon rank-sum, see Wilcoxon test Wilcoxon signed-ranks, see Wilcoxon test nonparametric test, 401, 405, 409, 519, 634 for three or more groups, 406–407 for two-way layout, 407–409 for two groups, paired data, 401 for unpaired (independent) samples, 404 nonrandom sampling, 37, 39, 48, 603 captive population, 48 convenience sample, 48 Delphi sample, 49 haphazard sampling, 49 P-values for, 603 quota sampling, 49 referred cases, 48 snowball sampling, 49 telephone sampling, 49 nonrespondent, 31, 33, 574, 575, 598 nonresponse, 10, 20, 29, 31, 34, 35, 48, 71, 95, 256, 290, 574, 575, 577 handling of, 20, 575, 577 nonsampling errors, 31, 256 nonsense correlation, 455, 605 nonsignificance, 300, 356, 519, 535, 607, 683 normal value, 174 population, 174 self, 174 normal distribution, see Gaussian normalization, 247, 578 normal level, see normal value normal range, 160, 165, 175, 176, 232, 559 normal value, 2, 169, 174, 200, 207, 616 implications of, 174 population, 174, 379 self, 174 Norway, 171, 244, 245 notations, 37 system of, 15 nuisance parameter, 381, 482, 522 null hypothesis, 277 number needed to treat, 355 CI for, 355 707 Index O obesity, measures of, 164 O’Brien–Fleming function, 302 observational study, 24, 51, 53, 63, 83 designs of, 51 reporting results of (STROBE), 65, 601 observed zero, 113 observer, 4, 56, 99, 130, 194, 392, 424, 458, 463, 495 blind, 95 bias, 33, 56, 94, 109, 573 variable competence of, 573 variation, 4, 9, 185, 193, 410 odds ratio, 209, 347, 357, 370, 631, 675 adjusted, 477, 482 confidence interval for, 361 in independent samples, 358 in matched pairs, 361 log of, 209, 362, 477 as a measure of strength of association, 489 pooled, 365 test of H0, 362 ogive, 123–124, 142, 174 one-sided alternative, 278, 314, 321, 360, 362, 368, 389, 603 one-sided bound, 265, 274, 324, 351, 455, 595, 603 one-sided test, see one-tailed test one-tailed test, 295, 302, 305, 323, 405 one-way, 67, 71, 73, 112–113, 313 ANOVA, 387, 389, 393, 396, 459 classification, 71, 112 design, 104, 613 experiment, 72 layout, 406 table, 113, 312, 313, 337, 339, 340 open trial, 89, 103 operations research, 21 opportunity, 76, 98 of bias, 90 of error, 213, 557, 558 of nonresponse, 31, 575 optimistic index, 571 OR, see odds ratio ordinal association, 492 measures of, 493 ordinal categories, 109, 138, 332, 469, 482, 483, 490, 497–498 association, 492 logistic, 484 partly, 109 trend in proportions, 332 ordinal data, see ordinal scale ordinal scale, 108, 109, 142, 311, 342, 424, 497 ordinary least squares, 430, 470 outcome, 3, 8, 18, 38, 51, 52, 53, 56, 82, 140, 181, 209, 270, 317, 322, 328, 348 adverse, 347, 350, 355, 356, 557 assessment, 84, 89, 90 measure, 58, 93, 168, 443, 537, 569 numeric, 110 prediction of, specification of, 90 variables, 19, 289, 537 outlier, 34, 120, 126, 127, 140, 144, 145, 174, 175, 307, 377, 390, 391, 401, 446, 448, 471, 486, 560, 571, 573, 591, 597 overanalysis, 597, 603 overfit (regression), 446 overmatching, 61 P pack-year, 166 paired sample, 270, 379; see also matched pairs pairwise comparison, 298, 335, 398–400 pairwise deletion, 446, 574 Palma measure of inequality, 243 parabolic curve, 454, 592 paradigm, 5, 37, 96, 212, 571 parallel group, 55 parameter, 17, 22, 30, 39, 53, 60, 108, 164, 177, 568, 604; see also population parameter biochemical, 531 biological, growth, 18 health, 2, 28, 242, 604 hematological, 562 logistic, 470, 473 medical, 2, 4, 18, 128, 604 nuisance, 381, 482, 522 pharmacokinetic, 78, 93, 594 physiological, 160, 460 regression, 426, 449, 671 statistical, 255, 263, 287, 381; see also population parameter survival, 515 validity, 101 parameter uncertainty, 5, 569 partial correlation, 14, 456 partitioning, 314, 336, 340 of chi-square, 315 of table, 314, 336 partogram, 133, 145, 150 path analysis, 535–536, 550 Pearson correlation, 453, 456; see also product–moment correlation PECOS, 370 pedigree chart, 153 percentile, 122, 123, 125, 149, 161, 176, 241, 614 percentile curve, 148, 161 percent of median, 161, 162 perinatal mortality rate, 229 perinatal mortality ratio, 229 period prevalence, 232 permutation test, 405 per-protocol analysis, 325 person-time, 234 months, 234, 238, 589 years, 234, 238, 504, 505, 506, 510, 580, 594 Peru, 224, 540 pharmacoepidemiology, 83 pharmacokinetic studies, 78, 594 pharmacokinetic parameters, 93 phases of a (clinical) trial, 81, 584 phase I, 39, 48, 81, 94 phase II, 82, 83, 97, 222, 229, 301 708 phase IIA, 82, 103 phase IIB, 82, 103 phase III, 83, 83, 97, 103, 222, 301 phase IV, see postmarketing surveillance phi coefficient, 490 physical quality of life index, 247 pie diagram, 136, 140, 143, 156, 615 exploded, 138 wedging, 138 Pillai trace, 538, 539 pilot study, 29, 30, 291, 297, 299, 446, 586 placebo, 32, 68, 74, 82, 89, 269, 288, 383, 578, 588 platykurtic, 172 plot, see graphs and diagrams residual, 442, 445, 624 pocket size health monitor, point estimate, 256, 264, 356, 380, 435 point prevalence, 232, 236 Poisson distribution, 258, 310, 311, 317, 482 Poisson regression, 486, 633 Poland, 230 polygon, frequency, 136 polynomial relation, 160 polynomial regression, 426, 632 polytomous, 307 categories, 108, 311, 316, 331, 488, 490, 492 covariates, 484 dependent, 469, 483, 542 regressor, 478 variable, 331, 335, 480, 523, 544 ponderal index, 160, 161, 165 Rohrer, 160 pooled chi-square, 365 pooled OR, 365 pooled RR, 365 pooled variance, 269, 382, 411, 414 population attributable fraction, 357, 373, 374 population attributable risk, 357 population normal, 174, 379 population parameter, 16, 41, 225, 257, 269, 287, 289, 291, 438 population pyramid, 221 population served per bed, 117, 244 per doctor, 121–122, 244 population (statistical), 37 concept of, 37 size, see size of population target, 9, 18, 22, 35, 45, 95, 189, 255, 290, 358, 426, 540, 568 population structure, 225 growth curve, 427 pyramid, 139, 221–222 stable, 223 standard, 225, 579 stationary, 223 United States, 225 Venezuela, 225 positive health, 159–160 positive predictivity, 4, 186, 187–190, 197, 262, 265, 571 positive skewed, see right-skewed posterior probability, 180, 182 Index postmarketing surveillance, 83 postneonatal mortality rate, 229, 230 poststratification, 83 posttest odds, 186 posttest probability, 191, 197, 606 potential years of life lost, 507 poverty index, 248 power analysis, 300 power (statistical), 280, 295, 301, 364, 368, 415, 585; see also statistical power power calculation, 281, 295, 300 power, discriminating, 540, 542 power transformation, 285 PP analysis, see per-protocol analysis P-P plot, 174 pragmatic trials, 92, 577 precision (degree of), 23, 39, 146, 290, 294, 423, 593 absolute, 257 relative, 257, 292, 368 predicted value, 391, 424, 428, 430, 437, 439, 441, 447 prediction interval, 437, 466 predictive model, 472, 475 predictive uncertainty, predictive validity, 187 predictivity, 186 negative, 186–187, 188, 189, 190, 197, 265, 474 positive, 4, 186–187, 197, 262, 265 and prevalence, 187 predictor, 209, 287, 425, 472, 601 pretesting, 30 pretest odds, 186 pretest probability, 186, 189, 191, 192, 197, 232, 606 prevalence rate, 133, 154, 178, 183, 187, 191, 197, 232, 569, 588 period, 232 point, 232 ratio, 63, 233 prevalence in relation to duration, 236 prevalence, meaning of, 188 prevalence study, 23 primary sampling unit, 43, 50 primordial factors, 102 principal component method, 547 principal components, 530, 547, 548 principles of experimentation, 68 principles of test of significance, 255 prior probability, 180, 182, 540 probability, 37, 177, 214, 263, 308, 469, 604 in assessment of prognosis, 180 binomial, 308, 309, 329, 402 in choice of treatment, 181 in clinical assessment, 177 conditional, 180, 348, 516 cumulative, 511, 516, 517 definition of, 177 of detecting a specified difference, 294 in diagnosis, 179–180 Gaussian, 259, 302 imprecise, 177 inverse, 180 joint, 178, 482 709 Index laws, 177 of misclassification, 539 multinomial, 316 personal, 180, 189, 197 posterior, 180, 182 posttest, 191, 197, 606 pretest, 186, 189, 191, 192, 197, 232, 606 prior, 180, 182, 540 of Type-I error, 146, 279, 281, 287, 288, 379, 398, 399, 412, 417, 432, 530 of Type-II error, 280, 559 probability proportional to size, 44, 47 probability sample, see sampling (methods) proband, 238 product–moment correlation, 452, 454, 456, 459, 530, 563–564 pro forma, 19, 61, 597 prognostic uncertainty, prognostic factors, 32, 53, 86, 99, 211, 308, 503, 522, 587 propensity score, 14, 487 prophylactic trial, 81 in clinic, 102 in field, 102 proportion, 34, 116, 595 confidence interval for, 263 inference from, 307 lower and upper bound, 265 in matched pairs, 328 in one sample, 309 trend in, 332 in 2×2 tables, 318 in two independent samples, 320 proportional hazards, 484, 522 how to check, 522 proportional hazard model, 484–485, 517, 522 proportional mortality, 231, 604 proportional reduction in error, 486, 491 proportionate sample, 42 proportion-by-probability plot, 174, 284 prospective design, see prospective study prospective study, 33, 53, 54, 55, 113, 297, 299, 318, 348, 353 biases in, 56 cohort study, 54 comparison group in, 55 demerits of, 58 historical, 54 longitudinal, 54 merits of, 57 performance of, 63 repeated measures, 55 selection of subjects, 55 variations of, 54 protective effect, 59, 349, 355, 371 protocol, 17, 20, 90, 99, 295, 573 contents of, 17 violation, 33, 325, 576 proximal measures, 240 pseudo, 545 R 2, 472, 473, 672, 687 sample, 570 Ptolemy, 278 P-value, 279, 402, 413 adjusted, 302 dramatic, 603 general procedure to obtain, 281 joint, 593 magic threshold, 602 misuse of, 602 nominal, 302 one-tail or two-tail, 603 for nonrandom sample, 603 publication bias, 34, 370 purchasing power parity dollars, 241 purposive sample, see sampling (methods) pyramid, population, 139, 221 pyramid, inverted, 26 Q Q-Q plot, 174, 284, 442, 648, 650, 663, 666 quadratic regression, 438, 443, 445, 446, 664 confidence band for, 438 qualitative, 159 agreement, 495 assessment, 111 categories, 447 characteristic, 159, 203, 319, 496 data, 111, 115, 484 dependent, 452, 469, 470 factors, 451 independent, 536, 539 measurement, 159, 460, 495 outcome, 209 predictor, 485 research, 13, 21 response, 108 variable, 111, 307, 335, 424, 448, 451, 460, 469, 488, 497, 537 quality, 111, 116, 215, 429, 556, 561, 595 of data, 572, 555 of decisions, of information, 29 of instrument, 208 of investigation, 96 of life, see quality of life of lots, 558 of measurement, 562 of medical care, 555, 557 of medical laboratory, 561 of research, of response, 29 of statistical model, 566 of (medical) test, 186 of trial, 100 quality control, 555, 561 chart, 142, 559, 561 of measurements, 20 in medical care, 555 in medical laboratories, 559 quality of life, 89, 168, 247, 531, 562 assessment, 168 index, 247 710 measuring, 169, 562 questionnaire, 169 quantile, 117, 122 in grouped data, 123 interpretation of, 124 from ogive, 124 by-quantile plot, 174, 284 in ungrouped data, 123 quantitative, 6, 27, 111, 159, 423, 452, 460, 470, 494, 531, 597 agreement, 460, 463, 495 analysis of codes, 597 aspects of medicine, 159 assays, 205 assessment, 203 characteristic, 70 covariate, 451, 522 data, 33, 111, 117, 276, 281, 290, 322, 377, 484, 486 dependent, 430, 452, 470, 486, 536, 539 independent, 531 measurement, 111, 159, 174, 203, 263, 460, 544, 592, 597 outcome, 53, 75, 398, 423 predictor, 485 relationship, 542 regression, 442, 469, 470, 471, 472, 473, 478, 482, 484, 487, 524, 532 response, 539 test, 159, 185 variable, 307, 361, 387, 392, 424, 432, 433, 439, 442, 456, 485, 488, 537, 544 quartile, 122, 123, 125, 132, 144, 174, 199, 570 quartile plot, 173, 284, 442 quasi-experiments, 86 quasi-random allocation, 88 questionnaire, 28, 30, 61, 189, 563 anonymous, 48 close-ended, 29 length of, 29 mailed, 575, 584 open-ended, 29 self-administered, 210 Quetlet index, 165 quintile, 122, 123, 243 quota sampling, 49 R radar graph, 150 radix, 504, 505 Ramanujan, 246 random allocation, 68, 71, 86, 87, 389, 390, 487, 583; see also randomization methods of, 87 quasi-, 88 random effect, 397 random error, 31, 36, 69, 461, 463, 559, 611 randomization, 8, 52, 68, 86 block, 88 cluster, 88 simple, 88 stratified, 88 randomization.com, 69, 88 Index randomized block experiment, 72 randomized controlled trial, 32, 82, 83, 150, 212, 305, 356, 576 failure of, 585 selection of participants for, 83 random numbers, 40, 69, 83, 87 random point, 44, 585 random sampling, 40 choice of method, 46 visual illustration, 46 random sampling method, 46 area, 47 cluster, 43 consecutive, 48 inverse, 47 multistage, 43 probability proportional to size, 44 sequential, 48 simple, 40 stratified, 41 systematic, 45 random selection, 39, 49, 60, 83, 86, 99, 463 random subgroup, 56 range, 120, 126, 166, 297, 433, 447, 491, 549 interquartile, 128, 144, 370, 594 normal, 165, 175, 176, 232, 559 reference, 604 semi-interquartile, 126 Studentized, 399 tolerance, 559 range check, 573 rank correlation, 455, 457 rank data, 109 rapid assessment, 44, 147 Rasch analysis, 29 rate, 10, 55, 116, 154, 221, 238, 455 abortion, 228 adjusted, 578 adult literacy, 240 adult mortality, 230 attack, 148, 238, 588 attrition, 48 birth, 146, 219, 222 case–fatality, 231, 237, 557 child mortality, 229 compliance, 588 complication, 273, 305, 598 conception, 517 cure, 288, 555 death, see death rate detection, 102 divorce, 245 error, 82, 183, 398, 557, 561, 597 exposure, 369 false negative, 188 false positive, 188 failure, 263, 273, 516 fertility, 230 general fertility, 10, 219 gross fatality, 581 growth, 163 Index hazard, 348, 484, 525 incidence, 116, 233, 234, 237, 318, 347 infant mortality, 10, 121, 142, 223, 229, 392, 433, 506 infection, 238 literacy, 240 mortality, 148, 230 neonatal mortality, 204, 229, 251, 617 net fatality, 557 perinatal mortality, 229 postneonatal mortality, 229 prevalence, 10, 133, 147, 188, 233 proportional mortality, 231, 604, 617 relief, 288 response, 270, 303, 326, 587, 598 secondary attack, 148, 238 standardized; see also standardized death rate stillbirth, 229 survival, 58, 111, 308, 309, 410, 510, 513 true negative, 188 true positive, 188 under-5 mortality, 154, 229 ratio, 39, 116, 312, 544 abortion, 228 albumin–globulin, 117 bed–population, see population served per bed birth weight, 160 comparative mortality, 227 cross-product, 358; see also odds ratio dependency, 242 doctor–population; see also population served per doctor enrollment, 240 F, 389; see also F-test hazard, 348, 372, 484, 518, 524, 687 incidence rate, 234 likelihood, 186, 472, 480, 687 maternal mortality, 230 odds, see odds ratio PAA, 387, 389 prevalence, 352, 588 prevalence rate, 63, 233, 588 perinatal mortality, 229 risk, 350, 352; see also relative risk sex, 203, 223, 605 skinfold, 178, 179 standardized mortality, 227, 579 stillbirth, 228 success, 92, 93, 487 variance, 388; see also F-test variation, 128 waist–hip, 109, 117, 145, 165, 204, 423, 479, 530 WBC to RBC, 117 ratio scale, 108, 131 R-code, 14, 643 RCT, see randomized controlled trial recalls lapse, 60, 65, 242 record linkage, 28 reestimation of sample size, 302 reference category, 478, 479, 481, 684 reference value, see normal value registration of births and deaths, 28 711 registration of trials, 100 regression, 423, see more details in the subsequent entries for regression choice of form of, 443 missing values in, 446 outliers in, 446 sum of squares, 440 regression assumptions, validity of, 430, 442 regression coefficient, 210, 426, 429, 430, 433, 439, 441, 447 adjusted, 439 CI on, 437 SE of, 437 significance of, 433, 441 standardized, 447, 449 unadjusted, 439 regression, adequacy of, 424, 440, 443 regression fitting, general method of, 429 regression, implications, of, 447 regression model, 430, 432, 442, 447, 566 regression (types of) asymptotic, 428 Cox, see Cox regression curvilinear, 425, 426, 427 explanatory, 429 hierarchical, 450 intrinsically linear, 426, 427, 445 linear, 425, 426, 428, 429, 432, 434, 443 logistic, see logistic regression mixed, 450 multilevel, 449, 451 multiple, 209, 425, 439, 442, 548, 574, 605 multiple linear, 426, 439, 424, 438–440, 443; see also multiple linear regression multivariate, 425, 538 multivariate multiple, 529, 532, 535 nonlinear, 427, 429, 443, 662 polynomial, 426 predictive, 429 quadratic, 438, 445, 446 ridge, 449 simple, 425, 436, 442, 574 simple linear, 426, 428, 433–435, 436, 437, 440, 443; see also simple linear regression spline, 450 through origin, 428, 441 univariate, 486, 532, 535 regression tree, 469, 485, 632 regressor, 425, 426, 429 choice of, 429, 441, 443, 447, 472 selection of, 430, 433 relationship, cause–effect (type of), 8, 18, 26, 51, 65, 178, 349, 455, 555, 605 versus correlation, 455, 604 criteria for, 605 relative risk, 93, 234, 265, 348 confidence interval for, 351 in independent samples, 348 limitation of, 607 in matched pairs, 353 pooled, 364 test of H0, 352 712 relative risk reduction, 356 reliability, 60, 120, 175 of conclusions, 20, 31, 57, 84, 87, 256, 410, 446 of devices, 4, 13, 22, 563, 564 of estimates, 39, 44, 441, 471 of instruments, see reliability, of devices interrater, 458, 495, 563 measure of, 256, 564 of measurements, 35, 564 of method, 460, 463 of results, see reliability, of conclusions of scoring systems, 211 test–retest, 565 of tools, see reliability, of devices repeatability, 190, 211, 565, 570; see also reliability repeated measurements, 130 repeated measures, 104, 387, 397 ANOVA, 397, 398, 421, 539, 633, 652 design, 74, 76 dichotomy in, 334 MANOVA, 539 study, 55, 66, 299, 486 replication, 69, 397 reproducibility, 488, 495, 563, 570; see also reliability reproduction rate, 221 gross, 221 net, 221 resampling, 570 bootstrap, 571 jackknife, 571 optimistic index, 571 research, credibility of, 22 research (types of), see medical research (types of) research synthesis, 9, 21; see also meta-analysis residual plot, 442, 445, 624 residual, 391, 429, 430, 435, 440, 442, 648 concept of, 428 heterogeneity, 68 independence of, 392, 456 forms of, 443 plot, 442, 445, 624 standardized, 650 sum of squares, 440 variance, 388 response, 7, 28 incomplete, 10 rate, 270, 303, 587, 598 surface, 145, 154, 438 variable, 70, 384, 388, 389, 394, 401, 425, 447 retrospective cohort, 54, 58, 66, 294 retrospective design, see retrospective study retrospective follow-up, 54 retrospective study, 58, 66, 111, 201, 252, 318, 359 performance of, 63 sampling methods in, 60 reverse causation, 63 ridge regression, 449 right-skewed, 171, 172, 173, 274 risk difference, 353, 357; see also attributable risk risk factor, 25, 45, 53, 232, 350, 471, 523, 676 Index choice of, 568 control of, 232 hypothesized, 57, 60 identification of, 524 risk ratio, 350; see also relative risk road-to-health (chart), 149 robustness, 17, 69, 555, 566, 568 robust, 383, 391, 539, 566, 569 to assumptions, 539 to departures, 262, 391, 592 to reviews, 20 ROC curve, 14, 154, 192, 196, 211, 281, 475, 593, 606 area under, 195, 211, 594, 677; see also area under the curve optimal threshold, 194 partial area, 197 predictivity based, 197 sensitivity-specificity based, 192 Youden index, 194, 195 Rohrer ponderal index, 160 R 2, 440, 441, 446, 596 adjusted, 441 pseudo, 472, 473, 672, 687 S sample, 15, 31, 37 biased, 45, 583, 584 independent, 269, 319, 326, 348, 351, 353, 358, 360, 380, 404, 456 paired, 270, 353, 361, 380 problems with, 583 representative, 26, 39, 45, 95, 103, 524 unbiased, 584 unrepresentative sample, 95 sample size, 39, 60, 96, 289, 294 in adaptive clinical trials, 300 determination of (in estimation setup), 289 determination (in test of hypothesis setup), 294 for equivalence trials, 298 formulas for estimation, 293 formulas for testing, 296 general considerations in estimation setup, 294 general considerations in test setup, 296 general procedure in estimation setup, 291 general procedure in test setup, 295 inadequate, 585 for hazards, 524 for inference on RR and OR, 366 for logistic regression, 298 in negative trials, 412 nomograms and tables of, 299 problems with, 586 reestimation in adaptive designs, 302 for sensitivity and specificity, 292 for survival studies, 517 thumb rules, 299 sample survey, 11, 25, 289 sampling, 37 advantages of, 38 choice of method of, 46 concepts, 37 713 Index error, 3, 31, 197, 256 fluctuation, 3, 31, 37, 71, 171, 179, 235, 255, 263, 279, 286 fraction, 39 frame, 39, 40, 43, 46 interval, 45 limitations of, 37, 38 plan, 24, 290 technique, see sampling (methods) terms used in, 38 unit, 30, 38, 39, 43, 47, 76 weight, 39 sampling distribution, 255, 284, 376 of p, 258 of x, 285 sampling (methods), 40 area, 47 circular, 45 cluster random, 43, 44, 47, 62 consecutive, 48 convenience, 48 Delphi, 49 haphazard, 49 inverse, 47 multistage random, 43, 45, 46 other random, 46 other nonrandom, 49 nonrandom, 39, 48, 60, 383, 629 probability, 39, 40 probability proportional to size, 44, 47 proportionate, 42 purposive, 48, 49 quota, 49 random, 46, 47, 140 referred cases, 48 self-weighting, 42 sequential, 48 simple random, 40, 256 snowball, 49 stratified random, 41, 42 systematic random, 45 telephone, 48 volunteers, 48 30×7, 44 San Francisco, 24, 601, 603 Sao Leopoldo, 45 saturated model (log-linear), 341 scale, choice of, 143, 601 scale (medical), Berg balance, 563 Expanded Disability Status, 209 Glasgow coma, 206 Hamilton depression, 93 visual analogue, 5, 203, 205, 208, 401 Yale observation, 206 scale (statistical), 33, 107 continuous, 94, 423, 594 discrete, 141, 424 interval, 108, 131 Likert, 29, 30, 1721 log, 78, 143, 203, 351, 371, 401, 522 metric, 108, 134, 317, 377, 424, 432 nominal, 107, 112, 129, 208, 311 numeric, 206 ordinal, 108, 311, 337, 424 rating, 29 ratio, 108, 131 semiordered, 109 Scandinavian, 227 scatter, 117, 125, 126, 426, 429, 443, 445, 454, 590 diagram, 140, 426, 430, 442, 453, 543, 591 plot, see scatter diagram plot matrix, 534, 657 schedule, 23, 562 schematic chart, 52, 125, 127, 133, 152, 548 score, factor, 550 scores (medical), 204 ability, 209, 563 APACHE, 206, 207, 424, 449, 482, 522, 678 Apgar, 161, 205, 401, 530 ASA, 206 Bishop, 192 bronchiolitis, 414 dosage, 333 Gleason, 146 Hacettepe, 206 heredity, 535 Likert, 30 McGill pain, 205, 568 nursing intervention, 211 pain, 53, 377, 401, 452, 456 paraspinal EMG, 101 paraspinal filtration, 101 peritonitis severity, 206 severity, 569 short and graphic ability, 209 visual acuity, 586 score (statistical), 29, 107, 203, 204, 209 discriminant, 540 factor, 550 geometric, 209 grading, 124 linear, 209, 332, 478 nonlinear, 333 risk, 207 Z, 149, 161, 162, 170, 256, 447, 464, 544, 578 scoring, method of, 208, 210 for diagnosis, 209 for graded characteristics, 208 scoring system, 3, 108, 203, 206, 210, 217, 485 for BRCA 21 mutation, 211 for Chagas disease, 207 for diagnosis, 205 for gradation of severity, 206 for Hirschsprung’s disease, 209 for multiple sclerosis, 209 for oral submucous fibrosis, 210 for prostate cancer, 211 reliability of, 211 for rickets, 211 for socioeconomic status, 241 714 for therapeutic intervention, 211 for thyroidism, 205 for unstable angina, 209 validity of, 210 screening test, 187, 189 SE, see standard error secondary attack rate, 148, 238 selective reporting, 100, 601 self-normal, 174 self-weighting, 42, 44, 47 semi-interquartile range, 126, 132 semiparametric model, 517 Senegal, 150 sensitivity, 101, 184, 194, 211, 292, 481, 562, 644 features of, 185 sensitivity analysis, 568 sensitivity experiment, 94 sequential sampling, 48 sequential trial, 95 serial correlation, 392, 442, 456 sex ratio, 203, 223, 605 sexual maturity rating, 165 Shapiro–Wilk test, 174, 285, 594 Sierra Leone, 230, 507 sign test, 401, 402 significance level, 279, 305 significance, medical, 12, 281, 286, 287, 321, 410, 606 versus statistical significance, 287, 410, 606 significance, statistical, see statistical significance significant digits, 599 similarity, measures of, 543 simple linear regression, 426, 428, 433–435, 440, 443, 466, 596 CI for parameters of, 435 confidence band for, 437 intercept and slope in, 434 test of hypothesis for the parameters of, 436 simple randomization, 88 simple random sampling, 40, 256 Singapore, 241, 605 single blind, 89 single linkage, 544 six sigma, 561 size of population, 42, 47, 442 size of sample, see sample size skewed distribution, 172, 388, 401, 510, 521, 594, 616 left, 172, 274 negative, see skewed distribution, left positive, see skewed distribution, right right, 171, 274, 517, 592, 594 skewness, coefficient of, 172 slope, 433–434, 451, 448, 455, 466, 601, 624, 668 confidence interval for, 435 test of hypothesis, 436 in regression, 78, 433, 434, 435, 448 standard error, 435 smoke pipe distribution, 171 smoking index, 167, 172, 206 snowball sampling, 49 Snow, John, 67 social classification, 241, 242, 335 Index social health, 224, 242 indicators of, 224, 240, 242 soft data, 597 versus hard data, 597 Somer d, 493 sources of uncertainty, 2, 6, 8, 77, 410 South Africa, 139, 243, 605 Spain, 235 Spanish Diabetes Association, 235 Spearman rank correlation, 455, 457, 467 Spearman–Brown coefficient, 564 specificity, 65, 184, 187 features of, 185 spectrum, 38, 84, 108, 231, 287, 547, 571, 597 of death, 231 of disease, 138, 185, 228 of etiologic factors, 61 of health, 240, 594 of health care, of subjects, 32, 44, 60, 571 of uncertainty, 2, 3, 4, of values, 15, 110 sphericity, 397, 398, 539, 653 spline, 207, 479 function, 93, 592 regression, 424, 449, 450 split-half consistency, 563 spot map, 154 spurious correlation, 455 spurious trend, 442 S-shaped curve, 470 stable population, 223 standard deviate, 259, 262, 284, 342 standard deviation, 40, 53, 126, 138, 162, 312, 378, 586, 652 in grouped data, 128 in population, 126 in sample, 126 in ungrouped data, 126 standard error, 39, 256, 320, 378, 429, 515, 557, 599, 606 of AR in independent samples, see standard error, of difference in proportions in independent samples of AR matched pairs, 354 of difference in means, 269 of difference in proportions in independent samples, 269 of difference in proportions in matched pairs, 271 of intercept in simple linear regression, 435 of kurtosis, 284 of lnOR in independent samples, 360 of lnOR in matched pairs, 362 of lnRR in independent samples, 351 of mean of y (in regression), 437 of median, 268 of number of errors, 557 of p, 256 of predicted y (in regression), 437 of regression coefficient, 435 of skewness, 284 of x, 256 of slope in simple linear regression, 435 of survival rate, 515 Index standard Gaussian distribution, 259, 282, 291, 481 standard Gaussian variate, 309 standardization, 4, 210, 224, 227, 459 for calculating adjusted rate, 578 of definitions, 572, 582 direct, 225, 227 indirect, 225, 227 of laboratories, 459, 559 lack of, 572, 578 of means, 578 methods, 578, 606 of rates; see also death rate of regression coefficients, 210, 446, 447 of values, 544, 578 standardized coefficient (in regression), 447 standardized death rate, 225, 227 direct, 225, 226 indirect, 225, 226 standardized deviate, see standard deviate standardized mortality ratio, 227, 579 standard normal distribution, see standard Gaussian distribution standard population, 225, 579 stationary population, 223 statistical analysis, 9, 53, 60, 70, 95, 110, 358, 567, 607 problems with, 594 statistical fallacy, 366, 583, 606, 608 statistical index, 20; see also index statistical inference, 9, 49, 86, 110, 255, 283, 307, 366, 383, 410, 576 statistical medicine, 205 statistical method, 1, 8, 31, 98, 124, 263, 307, 351, 387, 446, 530, 543 applicable to primary research, 21 can be annoying, 446 conventional, 287 to derive classification criteria, 539 that discovers natural groups, 543 to encourage critical and logical thinking, 13 to evaluate probability, 277 guide chart for, 15 handle errors, 557, 561 in health planning, helps in conclusion based on one sample, 255 heuristic explanations for, 35 in improving quality of medical care, 556 inapplicable, 595 inappropriate use, 602 intricate, 583 learning curve of, 583 more developed for quantitative data, 484 most assume Gaussian, 172 for managing uncertainties, 11 notorious for discovering significance, 287 objective of, 255 potent tool, 71 reduce the specter of uncertainty, 451 required to analyze paired data, 87 require random sample, 603 seem to belong to everyone, 608 to study variation, 107 vital role in empirical inferences, 608 715 valid conclusion for small samples, 271 validity of, 69 wrong, slow to surface, 583 statistical model, 248, 297, 566, 568; see also model limitations of, 566 quality of, 566 robustness of, 566 validity of, 567 statistical power, 13, 71, 280, 294, 368, 415 statistical significance, 30, 174, 255, 276 debate on, 286 and medical equivalence, 324 versus medical significance, 286, 287, 289 nature of, 410 philosophy behind, 276 subtleties of, 283 test of, 281, 287, 321 statistical test, 276, 278, 285, 383, 530, 586; see also tests of H0 versus CI, 286 philosophical basis of, 278 statistics (plural), 2, 66, 329, 375, 729, 811 statistics (singular), 16, 63, 230, 404, 771, 811, 837, 852 steady state, 75 stem-and-leaf plot, 134, 173, 284 stepwise methods, 210, 431, 474, 540, 657 backward elimination, 431, 441, 474 backward stepping, 432 forward selection, 431, 441, 474 stillbirth rate, 229 stillbirth ratio, 228 stochastic variable, 111, 425, 453, 530, 540, 628 stopping (early), 96, 301 Lan–deMets procedure, 301 O’Brien–Fleming procedure, 302 rule, 96, 301, 303 strata, 41, 42, 70, 364, 365, 475, 482, 590, 612, 676 stratification, 41,60, 364, 366, 487 stratified analysis, 35, 78, 364, 423; see also Mantel–Haenszel procedure inadequate, 366 stratified randomization, 88 stratified sampling, 40, 41, 46, 290 structural equation model, 536 structural zero, 113, 342 Studentized range, 399; see also Tukey test Student t-distribution, see t-distribution Student t-test, 378, 383 correlation coefficient, 455 crossover designs, 384 difference in means in paired samples, 380 difference in means in unpaired samples, 380, 383 one-sample (comparison with a prespecified mean), 378 regression coefficient, 436 study (types of), 22 analytical, 24, 36, 300 case, 24 case–control, 15, 32, 58, 60, 252, 268, 299, 361, 374, 499, 599, 610 case–referent, 59 cohort, 54, 59, 66, 237, 294 cross-sectional, 62, 63, 64, 189 716 descriptive, 18, 23, 36, 289, 294 ecological, 53 epidemiological, 51, 584 experimental, 24, 583 intervention, 67, 99 longitudinal, 54, 66, 237, 486, 635; see also prospective study nested case–control, 51, 59 observational, 19, 24, 51, 64, 65, 79, 100, 487, 519, 604, 611 pilot, 30, 35, 291, 446, 586 prospective, 53, 54, 55, 58, 299, 318 repeated measures, 55, 66, 299 retrospective, 54, 58, 111, 318, 320, 359 study design, 22, 23, 64; see also study (types of) study population, 43, 225 subsample, 31, 47, 56, 509, 570, 571, 576 Sudan, 174 sum of squares, 390 between, 390, 436, 539 error, 390, 436, 440 regression, 440 residual, see sum of squares, error total, 390 Type I, 396 Type II, 396 Type III, 396, 652, 653 within, see sum of squares, error superiority trial, 92, 267, 298, 595 test, 323–324, 635 surface chart, 148 surrogate, 47, 165, 185, 230, 240, 352, 487 end point, 91 Surveillance Epidemiology and End Results (SEER), 517 survey, 10, 24, 43, 146, 338, 584 ad hoc, 28 community, 38, 47 cross-sectional, 236, 237, 319, 584 customer satisfaction, 561 exhaustive, 235 health, 566 health interview, 23 immunization, 42 mailed questionnaire, 584 opinion, 290 rapid, 146 telephone, 48 survival, 58, 138, 208, 308, 424, 448, 508, 542, 591 actuarial method; see also survival, life table method analysis, 348, 503, 503, 508, 517, 632, 677 bias, 32, 629 comparison of functions, 517, 519 comparison of rates, 518 Cox model for, see Cox regression curve, 508, 511, 515, 517, 522, 678 data, collection of, 509 data, nature of, 508 duration, 6, 13, 108, 120, 231, 236, 266, 503, 508, 511, 513, 521, 526, 540, 566, 6-7, 681 factors affecting, 521 function, 511, 513, 626, 678 Kaplan–Meier method, see Kaplan–Meier method Index life table method, see life table method log-rank test, 519 measures of, 510 period, see survival, duration probability of, 179, 206, 207, 308, 511, 521 product–limit method, 513 rate, 58, 111, 308, 503, 518 time, see survival, duration Sweden, 174, 224 symmetric distribution, 172 synoptic chart, 148 synthesis, 9, 91; see also meta-analysis of research, 21 systematic random sampling, 45 T t-distribution, 262, 266, 455 t-test, see Student t-test table (for statistical inference), chi-square, 639 CI for median, 275 F, 640 Friedman, 408 Gaussian, 637 Kruskal–Wallis, 406 proportion, confidence bound for, 273 rank correlation, 458 Studentized range, 641 Student t, 638 Wilcoxon rank-sum, 405 Wilcoxon signed-ranks, 403 tables (types of), contingency, 112, 314, 331, 337, 338, 342, 364, 490, 518 frequency, 111, 113 incomplete, 113 multiple response, 113 target beneficiaries, 11 target population, 18, 28, 37, 39, 255, 311, 349, 377, 426, 509, 540, 558 size of, 10 Tarone–Ware test, 519 telephone sampling, 48 tertile, 123, 124, 131, 179, 210, 614 test criterion, 282, 286, 312, 319, 353, 387, 402, 538 test (medical), 183, 194, 330, 606 combination of, 190 gains from, 191 in parallel, 190 in series, 190 validity of, 12, 159, 183 when can be avoided, 192 test–retest reliability, 565 tests of H0, 284, 286 association, see chi-square test attributable risk, see test of H0, equality of proportions correlation coefficient, 455 equivalence, 285, 307, 323, 415 equality of dispersion matrices, 542 equality of location (nonparametric) of three or more groups, 407 717 Index equality of location (nonparametric) of two groups, 401–405 equality of location (nonparametric) in two-way layout, 407 equality of means in crossover trials, 384 equality of means in three or more groups, 389 equality of means in two groups, 380–381 equality of mean with specified value in one group, 378–379 equality of proportions, 320 equality of proportions in crossover trials, 322–323, 327 equality of regression lines, 448 equality of sensitivity, specificity, 330 equality of variances, 381 equality of survival patterns, 519 Gaussian pattern, 285, 315 goodness-of-fit, 311 homogeneity of dispersion matrices, 538 homogeneity of frequencies, 319 homogeneity of variances (homoscedasticity), 389, 391, 592 independence of attributes, 319 linear trend in proportions, 333 log-linear model, 340–343 logistic coefficient, 472, 474 mean difference, 380 multiple comparisons, 398–400 multivariate, 530 odds ratio, 362 one-sample mean, 539–540 proportion, 307 proportion difference, 320 rank correlation, 458 relative risk, 351 regression coefficient, 441 regression slope, 436 serial correlation, 442 sphericity, 398 specified pattern in proportions, 311 trend in proportions, 333 test of significance, 286, 287, 296, 307, 459, 596, 648 test statistic, see test criterion test (statistical), 276, 278, 285, 307, 383 Anderson–Darling, 174, 285, 594 ANOVA F, 387, 389, 391, 396, 440, 531, 648 Barnard, 327 Bartlett, 319, 548 Bonferroni, 298, 331, 342, 399, 400, 481, 652 Box M, 398, 538 Breslow, 364, 519, 542 Brown–Forsythe, 392 chi-square, see chi-square test Cochran Q, 334 Dunnett, 398 Durbin–Watson, 392, 442 binomial exact, 329 Fisher exact, 326, 331, 336, 362, 413 Friedman, 407 F, see tests (statistical), ANOVA F Gaussian Z, see Z-test Hosmer–Lemeshow, 474, 475, 672 Hotelling T 2, 537 Kolmogorov–Smirnov, 174, 285, 401, 594 Kruskal–Wallis, 406 Levene, 269, 381, 391, 411, 442, 522 log-rank, 517, 519–521, 687 Mann–Whitney, 196, 404 Mantel–Haenszel, 364 Mauchly, 398, 539, 635, 654 McNemar, 328–329, 330, 334 McNemar-Bowker, 337 multinomial, 312, 316, 483, 593 nonparametric, see nonparametric test Pillai trace, 538, 539 Shapiro-Wilk, 174, 285, 594 sign, 401–402 Student t, see Student t-test Tarone–Ware, 519 Tukey, 399–400, 603, 622 Welch, 269, 381, 392, 522, 682 Wilcoxon rank-sum, see Wilcoxon test Wilcoxon signed-ranks, see Wilcoxon test Wilks Λ, 532, 535, 538 Z, see Z-test tests (types of), diagnostic, 6, 206,183, 188, 192, 206 screening, 187, 189 statistical, 34, 93, 276, 278, 280, 283, 285, 307, 383, 409, 413 Thailand, 150 thematic map, 154 theory of chaos, 567 therapeutic equivalence, 93, 415 therapeutic trials, see clinical trial therapeutic uncertainty, three-dimensional diagram, 144 3-in-1 weight monitoring chart, 163 3-sigma limits, 170 three-way tables, 337 assessment of association, 338 log-linear model for, 340 ties, 402, 406, 492, 493 time-dependent covariate, see Cox regression time plot, 142 time series, 424, 442, 456 tolerance limit, 382, 464, 555, 560 total fertility rate, 221 training dataset, 567 trend, 2, curvilinear, 425, 426, 427, 446 cyclic, 428 linear, 333 nonlinear, 427 trend analysis, 433 trend assessment, 433 trend in means, see regression trend in proportions, 331, 332 trial (types of) clinical, see clinical trial crossover, see crossover design diagnostic, 101 equivalence, 93, 267, 298, 325 field, 102, 103, 104, 299 with interim appraisal, 95 multicentric, 82, 98 718 negative, 34, 101, 412 n-of-1, 94 noninferiority, 92, 93, 105, 267, 298, 325, 326, 344, 595 pragmatic, 92, 577 prophylactic, 81, 84, 99, 101, 102 randomized controlled, see randomized controlled trial screening, 99, 102 sequential, 94, 96 superiority, 92, 325, 595 therapeutic, 81, 255 two-stage, 96, 301, 303 uncontrolled, 83, 106, 613 up-and-down, 94, 386 vaccine, 81, 103 trimmed mean, 649 triple blind, 89 trivariate distribution, 113 Tschuprow coefficient, 490 T-score, 163, 200, 578, 616 Tucson, 351 Tukey test, 399–400, 603, 622 twin studies, 86 two-phase sampling, 14, 59 2-sigma limits, 170 two-sided alternative, 278, 283, 363, 405 ±2SD limits, 170 two-stage design, 96, 301 two-stage regression, 449 two-stage sampling, 290 two-tailed tests, 313, 603 two-way ANOVA, 392–393, 395, 396 two-way classification, 112 two-way design, 71, 72, 73 two-way tables, 112, 338, 341 log-linear model for, 341 Type-I error, 279, 281, 282, 287, 301, 324 magic threshold, 602 in two-stage design, 96 Type-I sum of squares, 396, 451, 670 Type-II error, 280, 324, 412, 417 Type-II sum of squares, 397 Type-III sum of squares, 397, 653, 670 U UFO (unidentified flying object), 212 Uganda, 245 unadjusted, 575 analysis, 452 association, 364 mean, 579 OR, 482 prevalence, 628, 629 rate, 65 regression coefficient, 439 unbalanced design, 71, 393, 539 unbalanced sample, 47 unbiased, 39, 257 sample, 48, 583, 584 allocation, 87 Index estimate, 39, 257, 443 mechanism, 573 result, 84 sample, 11, 19, 71, 583, 584 unblind (study/trial), 20, 32, 91, 95, 97 uncertainty, 1, 2, 4, 7, 11, 19, 76, 86, 97, 177, 255, 390, 430, 569, 586 collective, 97, 98 contributors to, control of, 213, 281, 608 data, 276 empirical, 71 equal, 97 in clinical practice, gaps, 19 in health and disease, 2, 159 in health planning and evaluation, 2, in medical decisions, 177 in medical research, impact of, 8, 12, 46 interval, 570 management of, 1, 11, 159, 539 measurement of, 7, 12, 177 minimize the role of, minimizing the impact of, 8, 12, 566 priori, 84 sources of, 2, 4, 6, 77, 255, 382, 410 specter of, 203, 451 spectrum of, 2, 3, 4, underlying, 98, 567, 584 (due to) variation, uncertainties (types of), aleatory, 35, 568 diagnostic, epistemic, 5, 7, 35, 53, 91, 166, 205, 206, 212, 444, 566 model, 569 parameter, 5, 569 predictive, prognostic, therapeutic, uncertainty analysis, 567, 568 uncertainty principles, 97 under-5 mortality rate, 229 UNDP, 245 UNFPA, 230 unit, 38 of inquiry, 38, 43 primary, see primary sampling unit of sampling, 30, 38, 41, 43, 47 United Arab Emirates, 507 United Kingdom, 6, 161, 246, 336, 508 United States, 23,102, 116,148,152,153, 162, 207, 225, 226, 227, 231, 241, 244, 245, 293, 394, 480, 588, 604 univariate, 203, 204 analysis, 398, 412, 530, 537, 539, 607, 654 distribution, 113 measure, 203 parameter, 539 quantity, 205, 530 regression, 486, 532, 534 Index result, 530, 534, 537, 538 score, 206 setup, 292, 530, 531, 537 technique, 530 test of significance, 534 universe, 37, 542, 564 concept of, 37 proxy, 570 unstandardized coefficients, 447 unweighted calculations, 50, 612 unweighted mean, 42 up-and-down trials, 94, 95 analysis of, 386 V vaccine trial, 82, 101, 103, 588 validation, 530 external, 190, 447, 488, 535, 540, 567, 568 internal, 447, 567 peer, 20 of results, 20 sample, 211, 217, 568 of scores, 211 validity, 4, 28, 31, 32, 57, 62, 65, 160, 186, 240, 246, 256, 562, 613 of antecedents, 35 of assessment, 159, 212 of assumptions, 391, 430, 433, 442, 583 of calculations, 109, 189 of conclusions, 55, 62, 387, 390 of data, 576, 583, 631 of design, 302 of devices, 22 of estimate, 31 of evidence, 212 external, 20, 70, 86, 568, 613 of F (test), 387, 388 of indicators, 65, 189 of inference, 49, 397, 576 of information, 30, 214, 355, 573 inherent, 185, 195 of instruments, 562 internal, 20, 86, 567, 613 of investigations, 63 of K–M survival, 515 of log-rank test, 521 measures of, 350, 563, 593 of measurements, 22, 35, 183, 600 of medical tests, 159, 183, 187 of meta-analysis, 371 of methods, 69 of parameters, 101 predictive, 187 of results, 20, 36, 56, 79, 83, 446 of statistical methods, 69 of scoring system, 209, 210, 211 of study, 32, 290 of survey, 290 of target population, 35 of trial, 89, 92, 97, 105, 106, 613 719 validity (types of), 562 concurrent, 211, 562 construct, 563 content, 563 criterion, 89, 563 face, 210, 562 variability, see variation variability, sources of, variable (types of), 110 antecedent, 13 binary, 107, 307, 451, 469, 488, 494, 523 binomial, 309 bivariate, 113 categorical, 481, 679, 683 concomitant, 53 confounding, 457; see also confounder and confounding factors continuous, 109, 110, 113, 134, 169, 261, 313, 423, 469, 481, 503, 594, 601, 680 dependent, 140, 340, 425, 429, 431, 433, 438, 452, 469, 483, 529, 543 deterministic, 111 dichotomous, 318, 495, 503; see also variable (types of), binary discrete, 110, 118, 262 dummy, 448, 452, 481,523 explanatory, 425, 429, 431, 480; see also variable (types of), independent independent, 423, 425, 431, 441, 452, 472, 524, 532, 539 indicator, 439, 448, 451, 478, 480, 522 metric, see scale (statistical) nominal, see scale (statistical) ordinal, see scale (statistical) outcome, 19, 74, 83, 289, 537, 567, 603, 649 Poisson, 311 polytomous, 331, 335, 480, 523, 544 predictor, 428; see also variable (types of), independent qualitative, 111, 258, 307, 336, 637, 488 quantitative, 111, 258, 387, 392, 432, 439, 494 random, 586 regressor, 426, 429, 448, 469, 485; see also variable (types of), independent response, 70, 322, 338, 384, 388, 401, 425; see also variable (types of), independent stochastic, 111, 530 target, 290, 425, 536, 542, 543 variance, 126, 281 analysis of, see ANOVA between-groups, 387; see also sum of squares calculation of, 127 of difference, 128 error, 388, 689; see also variance, residual in grouped data, 128 homogeneity of, see homoscedasticity multivariate analysis of, see MANOVA pooled, 269, 382, 414 in population, 126, 381, 389 residual, 387; see also sum of squares in sample, 126 of sum, 128 total, 387, 459; see also sum of squares in ungrouped data, 126 within-groups, 387; see also sum of squares 720 variance components analysis, 397 variance ratio, 388 variance-stabilizing transformation, 445 variation, 1, 81, 98, 131, 140, 164, 257, 425, 453, 549 coefficient of, see coefficient, of variation graphical methods to represent, 133 linear, 454 numerical methods to represent, 107 qualitative, 129 random, 585 systematic, 387 variation between groups, 388 individuals, 39, 175, 279, 322, 382, 390 means, 266 studies, 372 variation ratio, 128–129 variation (sources of), 390, 436, 570, 671 aleatory, 569 assessment, 4, 611 behavior, biologic, chance, diurnal, 1, 2, 397 environmental, genetic, inherent, 9, 196 instrument, intrinsic, 2, 8, 16, 196, 388, 611 laboratory, natural, 16, 326, 569, 611 normal, 163 observer, 4, 9, 185, 410, 388 response, 72 treatment strategies, variations of designs, 95 healthy life expectancy, 507 prospective studies, 54 regression, 449 tau, 493 varimax rotation, 549 velocity, 163 growth, 163 height, 164, 423 weight, 164 Venezuela, 225, 244 Venn diagram, 154, 181 visual analog scale, 5, 203, 205, 208, 401 visual display (of data), 143, 154 visualization (of data), 133 volunteer, 32, 39, 48, 584 based studies, 32, 56 W waist–hip ratio, 40, 109, 117, 145, 165, 204, 479, 530, 562 Wald statistic, 472, 480, 524 washout period, 75, 76, 78, 94, 322, 384, 386 Index weekend effect, 585 Weibull distribution, 633 weighted, 206, 285, 396 calculations, 44, 50, 612 kappa, 497 least squares, 430, 470 mean, 42 scores, 210 weight-for-age, 148, 161, 164 weight-for-height, 161 wedging of pie, 138 Welch test, 381, 392, 398, 522 West Africa, 188 Western Cape, 139 Western diet, 549 Wilcoxon test, 401, 402, 403, 519 rank-sum, 196, 404 signed-ranks, 401 Wilks Λ, 532, 535, 538, 539 Wilson interval, 265 World Bank, 222, 230, 241, 507 World Health Organization, 44, 100, 146, 164, 226, 507 Eleven Countries Study, 169 Expert Committee, 146 quality of life questionnaire, 169 standard population, 225 weight-for-age chart, 149 Y Yale observation scale, 206 Yates correction for continuity, 320 years of life lost, 248, 503 potential, 507 yoga, 316, 652 Youden index, 194 Yule Q, 489 Z Z, 448 Gaussian, 169, 259, 260, 266, 282, 284, 295 standardized deviate, 342, 259, 262, 284 Z-score, 149, 161, 162, 170, 259, 447, 464, 544, 578 Z-test for AR, 353 for cell frequencies, 342 for difference in means, 416 for difference in proportions, 320 for lnOR, 360 for lnRR, 352 for mean, 378 for OR, matched pairs, 362 for proportion in one groups, 310 for Wilcoxon rank-sum, 405 for Wilcoxon signed-ranks, 403 for testing correlation, 456