Statistical Process Control for the FDA-Regulated Industry

Statistical Process Control for the FDA-Regulated Industry

Also available from ASQ Quality Press:

Quality Risk Management in the FDA-Regulated Industry

José Rodríguez-Pérez

The FDA and Worldwide Quality System Requirements Guidebook for Medical Devices, Second Edition

Amiram Daniel and Ed Kimmelman

CAPA for the FDA-Regulated Industry

José Rodríguez-Pérez

Development of FDA-Regulated Medical Products: A Translational Approach, Second Edition

Elaine Whitmore

Medical Device Design and Regulation

Carl T. DeMarco

The Quality Toolbox, Second Edition

Nancy R. Tague

The Certified Six Sigma Green Belt Handbook

Roderick A. Munro, Matthew J. Maio, Mohamed B. Nawaz, Govindarajan Ramu, and Daniel J. Zrymiak

The Certified Manager of Quality/Organizational Excellence Handbook, Third Edition

Russell T. Westcott, editor

The Certified Six Sigma Black Belt Handbook, Second Edition

T. M. Kubiak and Donald W. Benbow

The ASQ Auditing Handbook, Fourth Edition

J.P. Russell, editor

The Internal Auditing Pocket Guide: Preparing, Performing, Reporting, and Follow-Up, Second Edition

J.P. Russell

Root Cause Analysis: Simplified Tools and Techniques, Second Edition

Bjørn Andersen and Tom Fagerhaug

To request a complimentary catalog of ASQ Quality Press publications, call 800-248-1946, or visit our website at www.asq.org/quality-press.

Statistical Process Control for the FDA-Regulated Industry

Manuel E. Peña-Rodríguez

ASQ Quality Press

Milwaukee, Wisconsin

American Society for Quality, Quality Press, Milwaukee 53203

© 2013 by ASQ

All rights reserved. Published 2013

Library of Congress Cataloging-in-Publication Data

Pena-Rodriguez, Manuel E.

Statistical process control for the FDA-regulated industry / Manuel E. Pena-Rodriguez.

pages cm

Includes bibliographical references and index.

ISBN 978-0-87389-852-2 (hardcover : alk. paper)

1. Process control—Statistical methods. 2. Manufacturing processes—United States—Quality control. I. Title.

TS156.8.P45 2013

658.5072’7—dc23 2013003376

ISBN: 978-0-87389-852-2

No part of this book may be reproduced in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher.

Publisher: William A. Tony

Acquisitions Editor: Matt T. Meinholz

Project Editor: Paul Daniel O’Mara

Production Administrator: Randall Benson

ASQ Mission: The American Society for Quality advances individual, organizational, and community excellence worldwide through learning, quality improvement, and knowledge exchange.

Attention Bookstores, Wholesalers, Schools, and Corporations: ASQ Quality Press books, video, audio, and software are available at quantity discounts with bulk purchases for business, educational, or instructional use. For information, please contact ASQ Quality Press at 800-248-1946, or write to ASQ Quality Press, P.O. Box 3005, Milwaukee, WI 53201-3005.

To place orders or to request ASQ membership information, call 800-248-1946. Visit our website at http://www.asq.org/quality-press.

To my daughter Stacey Marie and to my mom Sonia Rodríguez. Thanks for always being my inspiration in everything I do.

And to my best friend José (Pepe) Rodríguez-Pérez. Thanks for believing in me and helping me to make this dream come true.

List of Figures and Tables

Figure 3.1 Concepts of process variation as compared to customer specifications.

Figure 4.1 Symbols used for some parameters and statistics.

Figure 4.2 Data collection matrix.

Figure 4.3 Sample size calculation—continuous data, example 1.

Figure 4.4 Sample size calculation—continuous data, example 2.

Figure 4.5 Sample size calculation—continuous data, example 3.

Figure 4.6 Sample size calculation—discrete data.

Figure 4.7 Histogram with descriptive statistics for the weight of a tablet.

Figure 4.8 Mode, median, and mean in a normal distribution.

Figure 4.9 Mode, median, and mean in a nonnormal distribution.

Figure 4.10 Histogram and descriptive statistics for nonnormal data.

Figure 5.1 Histogram for thread diameter.

Figure 5.2 Multiple histograms for thread diameter.

Figure 5.3 Box plot.

Figure 5.4 Multiple box plots.

Figure 5.5 Dot plot.

Figure 5.6 Multiple dot plots.

Figure 5.7 Defects Pareto diagram.

Table 5.1 Application of a weighting factor to the Pareto diagram.

Figure 5.8 Weighted Pareto diagram.

Figure 5.9 Weighted Pareto diagram with the other bar.

Figure 5.10 Scatter plot for tablet weight versus dissolution time.

Figure 5.11 Histogram and process performance indices for pin diameter.

Figure 5.12 Run chart for pin diameter.

Figure 5.13 Run chart for days to complete a laboratory investigation.

Figure 5.14 Run chart showing clusters.

Figure 5.15 Nonparametric run test showing clustering and nonrandomness of data.

Figure 5.16 Run chart showing mixtures.

Figure 5.17 Nonparametric run test showing mixtures and nonrandomness of data.

Figure 5.18 Run chart showing trends.

Figure 5.19 Nonparametric run test showing trends and randomness of data.

Figure 5.20 Run chart showing oscillations.

Figure 5.21 Nonparametric run test showing oscillations and randomness of data.

Figure 5.22 Normal and nonnormal data.

Figure 6.1 Gage R&R data collection matrix.

Figure 6.2 Percent contribution of each component.

Figure 6.3 Percent precision-to-tolerance and percent gage R&R.

Figure 6.4 Sources of variation in a measurement systems analysis.

Figure 7.1 Voice of the process versus voice of the customer.

Figure 7.2 Capable process.

Figure 7.3 Incapable process.

Figure 7.4 Process capability and process performance indices.

Figure 7.5 Interpretation of process capability and process performance indices.

Figure 7.6 Histogram and descriptive statistics for nonnormal data example.

Figure 7.7 Normal process capability analysis for nonnormal data example.

Figure 7.8 Box-Cox transformation process capability analysis for nonnormal data example.

Figure 7.9 Normality test for original data and Box-Cox transformed data.

Figure 7.10 Histogram and descriptive statistics for net weight.

Figure 7.11 Individuals and moving range chart for net weight.

Figure 7.12 Normal process capability analysis for net weight.

Figure 7.13 ImR chart for before and after analysis for net weight.

Figure 7.14 Process capability analysis for net weight after the improvement project.

Figure 8.1 Possible decisions in the acceptance or rejection of a lot.

Figure 8.2 One-sample t-test example.

Figure 8.3 Bartlett’s test.

Figure 8.4 Two-sample t-test.

Figure 8.5 Box plots for manufacturing cycle time comparison example.

Figure 8.6 ANOVA test for manufacturing cycle time comparison example.

Figure 8.7 Box plots for quality index comparison example.

Figure 8.8 ANOVA test for quality index comparison example.

Figure 8.9 Box plots for machine and material example.

Figure 8.10 Two-way ANOVA test for machine and supplier example.

Figure 8.11 One-sample sign test example.

Figure 8.12 Histogram and descriptive statistics for tablet hardness example.

Figure 8.13 Two-sample Mann-Whitney test for tablet hardness example.

Figure 8.14 Histogram and descriptive statistics for viscosity example.

Figure 8.15 Kruskal-Wallis test for viscosity example.

Figure 8.16 Box plots for transdermal patch adhesiveness example.

Figure 8.17 F-test for transdermal patch adhesiveness example.

Figure 8.18 Box plots for transdermal patch adhesiveness example.

Figure 8.19 Bartlett test for transdermal patch adhesiveness example.

Figure 8.20 Histogram and descriptive statistics for transdermal patch adhesiveness example.

Figure 8.21 Box plots for laboratory test evaluation example.

Figure 8.22 Levene test for laboratory test evaluation example.

Figure 9.1 Types of correlation.

Figure 9.2 The least squares method.

Figure 9.3 Perfect correlation.

Figure 9.4 Strong and weak correlation.

Figure 9.5 Histogram and descriptive statistics for residuals analysis.

Figure 9.6 Individuals control chart for residuals analysis.

Figure 9.7 Scatter plot for ease of communications versus customer satisfaction index.

Figure 9.8 Regression analysis for ease of communications versus customer satisfaction index.

Figure 9.9 Four-factor multiple regression analysis for customer satisfaction index.

Figure 9.10 Two-factor multiple regression analysis for customer satisfaction index.

Table 10.1 Design of experiments terminology.

Table 10.2 Relationship between number of levels and factors.

Table 10.3 Degree of fractionalization versus resolution.

Table 10.4 Data table for blocking experiment.

Figure 10.1 Results for blocking experiment.

Table 10.5 Repetition in DOE.

Table 10.6 Replication in DOE.

Table 10.7 Repetition and replication in DOE.

Table 10.8 Replicated full factorial design example for tablet hardness.

Figure 10.2 Main effects plots for tablet hardness example.

Figure 10.3 Interaction plots for tablet hardness example.

Figure 10.4 Factorial design analysis for tablet hardness example.

Figure 11.1 Tests for nonrandom patterns in control charting.

Figure 11.2 Individuals and moving range chart for pH.

Figure 11.3 X-bar and R chart for tablet weight.

Figure 11.4 X-bar and s chart for bottle weight.

Figure 11.5 Attributes chart example.

Table 11.1 Calculations for attributes chart example.

Figure 11.6 p-chart for percentage of manufacturing batch records with errors.

Figure 11.7 np-chart for number of manufacturing batch records with errors.

Figure 11.8 c-chart for number of errors per manufacturing batch record.

Figure 11.9 u-chart for average number of errors per batch record per week.

Figure 12.1 Individuals control chart for tablet hardness example.

Table A.1 Variable and attribute data applications.

Table B.1 Applications for graphical tools.

Table B.2 Applications for statistical tools.

Figure C.1 Graphical summary for Amiplinato tablets APR.

Figure C.2 Process capability analysis for Amiplinato tablets APR.

Figure C.3 Individuals and moving range chart for Amiplinato tablets APR.

Figure C.4 Two-year comparison for Amiplinato tablets APR.

Figure C.5 ANOVA test for Amiplinato tablets APR.

Figure C.6 Bartlett test for Amiplinato tablets APR.

Figure D.1 Most commonly used hypothesis tests.

Preface

Over the centuries, the quality of products and services has been one of the common characteristics of successful organizations. The term quality has evolved through the generations. Philosophies such as quality control, quality assurance, and total quality management have been recognized at different times. Nevertheless, all these philosophies share something in common: the use of statistical process control (SPC) to achieve higher levels of excellence. The concept of SPC applies to any type of industry: automotive, textiles, pharmaceutical, biologics, medical devices, electronics, aerospace, banking, educational services, and so on.

With the advances in technology, more people are immersed in the SPC arena every day. Computer software such as Minitab, Statgraphics, SigmaXL, and others, make the analysis of data a simpler task. However, most of the questions that people ask me every day are not about how to perform the analysis once the person determines which tool to use, but about which is the appropriate tool to use for each specific situation.

The focus of this book is to understand and apply the different SPC tools in a company regulated by the Food and Drug Administration (FDA): those that manufacture pharmaceutical products, biologics, medical devices, food, cosmetics, and so on. The book is not intended to provide an intensive course in statistics; instead, it is intended to provide a how-to guide about the application of the diverse array of statistical tools available to analyze and improve the processes in an organization regulated by FDA. This book is aimed at engineers, scientists, analysts, technicians, managers, supervisors, and all other professionals responsible to measure and improve the quality of their processes. Although the examples and case studies presented throughout the book are based on situations found in an organization regulated by FDA, the book can also be used to understand the application of those tools in any type of industry.

The book comprises 12 chapters and four appendixes. In Chapter 1, the regulatory importance of SPC is presented. Some of the FDA regulations and guidances are analyzed in terms of the agency’s expectations about the use of statistical process control tools. Also, some of the international standards applicable to the life sciences industry are analyzed for SPC requirements. Chapter 2 presents various instances in which FDA has issued observations about the misuse of SPC tools. Also, the concepts of SPC and corrective action and preventive action (CAPA) are integrated in this chapter.

Then, Chapter 3 presents the concept of process variation. The common causes and special causes of variation are explained in detail. Chapter 4 presents some basic statistical concepts, such as types of data, sampling, descriptive statistics, the normal distribution, and so on. Next, Chapter 5 presents some of the most useful graphical tools with which to start analyzing processes. Tools such as the histogram, dot plot, box plot, Pareto diagram, and others, as applied to several FDA-regulated industries, are presented in the chapter.

In Chapter 6, one of the most important but less frequently used tools is presented: the measurement systems analysis. In this chapter, the importance of addressing measurement system variability prior to implementing any other improvement initiative is thoroughly explored. Chapter 7 presents the concept of process capability. Here, we study the different indices used to measure capability: Cp, Cpk, Pp, and Ppk. Then, in Chapter 8, an introduction to hypothesis testing is presented. Several tools used to compare means, medians, and variances are introduced for normal and nonnormal data. Many examples are provided detailing the use of these tools in an FDA-regulated organization.

Chapter 9 explains how to use regression analysis to understand the relationship between input variables and output variables. Then, Chapter 10 provides a brief introduction to design of experiments and its application in an FDA-regulated environment. The concepts of full factorial and fractional factorial experiments are introduced in this chapter. In Chapter 11, control charts are introduced