Gynecologic Oncology E-Book

Steven A. Vasilev

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 1160

Veröffentlichungsjahr: 2011

Beschreibung

• New edition fully includes Grades and Levels of evidence, providing a better idea of what the sum total of existing evidence suggests regarding key topics
• Grades and levels of evidence consistently highlighted throughout
• Greater number of tables and algorithms ("decision trees") for key decision-making areas
• "Scope of concept" introduction to chapters are significantly enhanced, allowing the reader to reference the concepts while reviewing the specific details in the clinically oriented chapters
• The only available book with an evidence-based approach covering this specific subject matter in one compendium

Leseprobe

Contents

Cover

Title Page

Copyright

Dedication

Foreword to the Second Edition

Contributors

Part One: General Principles

Chapter 1: Introduction

TOWARD EVIDENCE-BASED MEDICAL PRACTICE

PERIOPERATIVE AND CRITICAL CARE ECONOMICS

OPERATIONAL TOOLS OVERVIEW

REFERENCES

Chapter 2: Evidence-Based Medicine and Decision Support

WHAT IS EVIDENCE-BASED MEDICINE?

EXTERNAL DATA GATHERING AND RELEVANCE INTERPRETATION TOOLS

WHAT IS THE “BEST” EVIDENCE?

EVALUATING PROSPECTIVE RCTS (EBM LEVEL I)

EVALUATING CONTROLLED PROSPECTIVE NONRANDOMIZED STUDIES (EBM LEVEL II-1)

EVALUATING OBSERVATIONAL STUDIES (EBM LEVELS II-2, II-3, AND III)

STATISTICS ISSUES: DESIGN, POWER, SAMPLE SIZE, AND CLINICAL UTILITY

DECISION ANALYSIS PRIMER

ECONOMIC ANALYSIS PRIMER

TOWARD EFFECTIVE GUIDELINE DEVELOPMENT AND IMPLEMENTATION

BEST PRACTICES LOOP MODEL

REFERENCES

Chapter 3: Vascular Access and Other Invasive Procedures

INTRODUCTION

VASCULAR ACCESS PROCEDURES

PLEURAL CAVITY PROCEDURES

ABDOMINAL CAVITY PROCEDURES

REFERENCES

Chapter 4: Fluids, Electrolytes, and Nutrition

FLUID AND ELECTROLYTE HOMEOSTASIS

ACID–BASE HOMEOSTASIS

FLUID AND ELECTROLYTE THERAPY

PARENTERAL AND ENTERAL NUTRITION

THE GLUCOSE SYSTEM

THE LIPID SYSTEM

THE THREE-IN-ONE SYSTEM

COMPLICATIONS OF PARENTERAL NUTRITION

NONNUTRITIONAL EFFECTS OF PARENTERAL NUTRITION

INDICATIONS FOR PARENTERAL NUTRITION

COMPARING METHODS OF TPN

SELECTING THE TPN REGIMEN

HOME PARENTERAL NUTRITION

COMPARISON OF ENTERAL AND PARENTERAL NUTRITION

REFERENCES

Part Two: Perioperative Management of Gynecologic Surgery

Chapter 5: Preoperative Evaluation

INTRODUCTION

THE IMPACT OF ROUTINE TESTING

LABORATORY TESTS

ELECTROCARDIOGRAM

CHEST RADIOGRAPH

PULMONARY FUNCTION TESTING

SUMMARY

REFERENCES

Chapter 6: Postoperative Surveillance and Perioperative Prophylaxis

PHYSIOLOGICAL RESPONSES TO SURGERY (LEVEL II-2)

GENERAL PERIOPERATIVE PROPHYLAXIS

GENERAL PROBLEMS IN POSTOPERATIVE MANAGEMENT

STOMAL CARE

SPECIAL MANAGEMENT CONSIDERATIONS IN THE HIGH RISK SURGICAL PATIENTS

REFERENCES

Chapter 7: Perioperative Infections: Prevention and Therapeutic Options

POSTOPERATIVE FEVER

SURGICAL INFECTION GENESIS

RISK FACTORS FOR INFECTION

SURGICAL PROCEDURE CLASSIFICATION

SURGICAL SITE INFECTIONS

SSI MANAGEMENT

PHARMACOECONOMICS

CLINICAL REALITIES

SSI SURVEILLANCE AND INFECTION CONTROL PROGRAM

REFERENCES

Chapter 8: Intraoperative and Perioperative Considerations in Laparoscopy

INTRODUCTION

PHYSIOLOGY OF PNEUMOPERITONEUM

COMPLICATIONS

LAPAROSCOPIC PROCEDURES DURING PREGNANCY

SIMULATION TRAINING

REFERENCES

Part Three: Oncologic Perioperative Decision Making

Chapter 9: Cervical Carcinoma

INTRODUCTION

THE SURGICAL TREATMENT OF CERVICAL CANCER

TECHNIQUES OF RADICAL HYSTERECTOMY

ROUTINE POSTOPERATIVE ASSESSMENT AND MANAGEMENT

SURGICAL COMPLICATIONS

LESS INVASIVE THERAPIES FOR EARLY STAGE CERVICAL CANCER

MANAGEMENT OF COEXISTING PREGNANCY AND CERVICAL CANCER

URETERAL OBSTRUCTION

REFERENCES

Chapter 10: Endometrial Cancer

INTRODUCTION

FAMILIAL GENETIC SYNDROMES

PRECURSOR LESIONS

PREOPERATIVE EVALUATION

STAGING

POSTOPERATIVE SURVEILLANCE

POSTOPERATIVE HORMONE REPLACEMENT

ADJUVANT THERAPY IN EARLY DISEASE

TREATMENT OF ADVANCED DISEASE

REFERENCES

Chapter 11: Pelvic Masses and Ovarian Carcinoma

INTRODUCTION

OVARIAN CANCER DIAGNOSIS UNCERTAIN

DIAGNOSTIC CONSIDERATIONS

OVARIAN CANCER DIAGNOSIS CERTAIN

REFERENCES

Chapter 12: Molar Gestation

INTRODUCTION

TROPHOBLASTIC EMBOLIZATION AND CARDIORESPIRATORY INSUFFICIENCY

PULMONARY EDEMA

ANEMIA

COAGULOPATHY

PREGNANCY-INDUCED HYPERTENSION/ECLAMPSIA

HYPERTHYROIDISM

HYPEREMESIS GRAVIDARUM

ENDOMYOMETRITIS

POSTEVACUATION DELAYED BLEEDING

MANAGEMENT GUIDELINES FOR MOLAR PREGNANCY EVACUATION

REFERENCES

ADDITIONAL POSSIBLE TOPICS FOR FUTURE EDITIONS

Chapter 13: Perioperative Issues in the Management of Vulvar Cancer

INTRODUCTION AND BACKGROUND

LESS RADICAL SURGERY

COMPLICATIONS OF THERAPY

VULVAR RECONSTRUCTION

RADIATION THERAPY

REFERENCES

Part Four: Complimentary Medicine/Supportive Care

Chapter 14: Perioperative Psychosocial Considerations

INTRODUCTION

SEXUAL/EMOTIONAL PROBLEMS

ISSUES OF DEATH AND DYING

PSYCHOLOGICAL COMPLICATIONS IN INTENSIVE CARE UNIT

SUMMARY

REFERENCES

Chapter 15: Pain Management in Gynecologic Oncology

INTRODUCTION

CLINICAL ASSESSMENT

PAIN ASSESSMENT

CHARACTERISTICS OF PAIN

COMMON CANCER PAIN SYNDROMES

MEASUREMENT OF PAIN

MANAGEMENT OF PAIN

MONITORING OF PAIN MANAGEMENT QUALITY

CONCLUSIONS

REFERENCES

Chapter 16: Fertility Preservation in the Gynecologic Cancer Patient

INTRODUCTION

RADIATION

CHEMOTHERAPY

PRESERVATION OF OVARIAN FUNCTION DURING CHEMOTHERAPY

PRESERVATION OF OVARIAN FUNCTION DURING RADIATION

CERVICAL CANCER

ENDOMETRIAL CANCER

OVARIAN CANCER

REFERENCES

Chapter 17: Perioperative Herbal and Supplement Use

EVIDENCE-BASED AND COMPLEMENTARY AND ALTERNATIVE MEDICINE

POTENTIALLY DETRIMENTAL PERIOPERATIVE SUBSTANCES

POTENTIALLY BENEFICIAL PERIOPERATIVE SUBSTANCES

EICOSAPENTAENOIC ACID

ORNITHINE KETOGLUTARATE

CONCLUSION

REFERENCES

Chapter 18: End-of-Life Decision Making

PAIN CONTROL

INTESTINAL OBSTRUCTION

NUTRITIONAL MANAGEMENT

ELECTROLYTE CHANGES

NAUSEA AND VOMITING

CONSTIPATION

ASCITES

RESPIRATORY FUNCTION

FATIGUE

BLEEDING

PACKED RED BLOOD CELLS

DECISION MAKING

EVIDENCE SUMMARY

REFERENCES

Index

Copyright © 2011 by John Wiley & Sons, Inc. 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, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone], or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations 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 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 [phone], outside the United States at [phone] or fax [phone].

Wiley also publishes 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:

Lentz, Scott E. Gynecologic oncology : evidence-based perioperative and supportive care / Scott E. Lentz, Allison E. Axtell, Steven A. Vasilev. – 2nd ed. p. ; cm. Rev. ed. of: Perioperative and supportive care in gynecologic oncology: evidence-based management / edited by Steven A. Vasilev. c2000. Includes bibliographical references. ISBN 978-0-470-08340-6 (cloth) 1. Generative organs, Female–Surgery. 2. Evidence-based medicine. 3. Clinical medicine–Decision making. I. Axtell, Allison E. II. Vasilev, Steven A. III. Perioperative and supportive care in gynecologic oncology. IV. Title. [DNLM: 1. Genital Neoplasms, Female–surgery. 2. Critical Care–methods. 3. Evidence-Based Medicine. 4. Gynecology–methods. 5. Perioperative Care–methods. WP 145] RG104.P393 2010 618.1′059–dc22

2010033322

This manuscript is humbly offered as an attempt to focus our lives as physicians onto our greatest goal–superior quality care delivered in a compassionate and meaningful way.

The product of this effort would not have been possible without our richest sources of support. For SEL–this is for my children, John, Madison and Timothy, who continually remind me that life is what happens while I'm waiting for the next big thing; and my wife Amelia, a savior in more ways than I can verbalize. For SAV–this is for my mother, Katharina, a lifelong source of support and encouragement; my two sons, Alex and Andrei, sources of emotional inspiration and vibrant reflections of a continuous youthful zeal for learning, and my loving wife Joyce who regularly redirects me towards those people, things and moments that matter most in life.

Most importantly this is for our patients, an endless source of fascination, learning, instruction and inspiration.

Scott E. Lentz, MD Allison E. Axtell, MD Steven A. Vasilev, MD, MBA

Foreword to the Second Edition

The information provided in this edition is much needed in the everyday management of patients with gynecologic cancer. This clinically practical textbook provides guidelines that are critical for excellent preoperative, perioperative, and supportive care of gynecologic oncology patients. Special effort has been made to employ the principles of evidence-based medicine derived from scientific studies, whenever available. A separate discussion on why evidence-based medical treatment is so important is included.

The second edition differs from the first one in several important ways: In this edition, special attention has been paid to evidence-based reviews of each of the major disease sites in gynecologic oncology. Not included in this edition are ancillary topics of the first edition, such as intensive care management, anesthetic issues, and perioperative radiology. These topics, though very important, were felt to distract from the specific goals in the title. A very practical chapter on the prevention and treatment of laparoscopic complications is included. A new, but very useful, section on the management of end-of-life complications (common in patients with advanced gynecologic cancer) such as tumor-related bowel obstruction, persistent ascites, and pleural effusion is also included. In addition, this edition has discussions on the ever-challenging pain management issues faced by the physicians and caregivers at the end of life of the patient. The role of perioperative herbal supplements, along with necessary precautions, is discussed in detail.

Those who perform surgeries for gynecologic cancer as well as those who carry out benign gynecologic procedures will find this edition to be a most helpful guide. Residents and fellows in obstetrics and gynecology will find an abundance of useful information in this text.

Leo D. Lagasse, MD Professor Emeritus Department of Obstetrics and Gynecology David Geffen School of Medicine at UCLA

Contributors

Denise Aberle, MD, Professor, Department of Radiological Sciences, UCLA School of Medicine, Los Angeles, CA

Malaika E. Amneus, MD, Assistant Professor, Division of Gynecologic Oncology, Olive View-UCLA Medical Center, Sylmar, CA

Allison E. Axtell, MD, Gynecologic Oncologist, Division of Gynecologic Oncology, Southern California Permanente Medical Group, Los Angeles Medical Center, Los Angeles, CA

Margarett C. Ellison, MD, MHA, Gynecologic Oncologist, Piedmont Gynecologic Oncology, Atlanta, GA

Nicole Fleming, MD, Fellow, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, UCLA Geffen School of Medicine, Los Angeles, CA

Laszlo Z. Galffy, MD, Anesthesia and Pain Management Center of Glendale, CA

Paul Koonings, MD, Director of Gynecologic Oncology, Kaiser Permanente, San Diego, CA

Leo Lagasse, MD, FACS, FACOG, Director of Gynecologic Oncology, Cedars-Sinai Comprehensive Cancer Center, Professor of Gynecology, University of California, Los Angeles, CA

Scott E. Lentz, MD, Regional Director of Gynecologic Oncology Services, Southern California Permanente Medical Group, Kaiser Permanente Los Angeles Medical Center, Los Angeles, CA

Kathryn F. McGonigle, MD, FACS, FACOG, Assistant Professor, UCLA School of Medicine, Division of Gynecologic Oncology, UCLA School of Medicine Los Angeles, CA

Judith McKay, PhD, Psychologist

R. Wendel Naumann, MD, Director Minimally Invasive Surgery in Gynecologic Oncology, Associate Director of Gynecologic Oncology, Blumenthal Cancer Center, Carolinas Medical Center, Charlotte, NC

Matthew Powers, MD, Resident, Department of Surgery, Keck School of Medicine, University of Southern California, Los Angeles, CA

Howard Silberman, MD, FACS, Professor of Surgery, Keck School of Medicine, University of Southern California, Los Angeles, CA

Harriet Smith, MD, Professor, Division of Gynecologic Oncology, Montifiore Medical Center, Bronx, NY

Amy Stenson, MD, Assistant Professor, Department of Obstetrics and Gynecology, UCLA Geffen School of Medicine, Los Angeles, CA

Devansu Tewari, MD, Gynecologic Oncologist, Southern California Permanente Medical Group, Sand Canyon Hospital, Anaheim, CA

Fidel A. Valea, MD, Associate Professor, Division of Gynecology Oncology, Department of Obstetrics and Gynecology, Duke University Health System, Durham, NC

Clayton A. Varga, MD, MHSM, Pasadena Rehabilitation Institute, Pasadena, CA

Alexander Vasilev, BA, AB, Clinical Research Associate, Department of Obstetrics and Gynecology, Kaiser Permanente Los Angeles Medical Center, Los Angeles, CA

Steven A. Vasilev, MD, MBA, FACOG, FACS, Director, Surgical and Radiation Oncology Clinical Trials Director, Gynecologic Oncology Integrative Therapies, Kaiser Permanente Los Angeles Medical Center, Clinical Professor, UCLA David Geffen School of Medicine, Los Angeles, CA

Part One

General Principles

Chapter 1

Introduction

Steven A. Vasilev, MD, MBA and Scott E. Lentz, MD

“Learning without thinking is useless. Thinking without learning is dangerous.”

–Confucius

TOWARD EVIDENCE-BASED MEDICAL PRACTICE

The very roots of Osler's apprentice-based medical education and practice, to which we still largely adhere, are being severed. Lest this be interpreted as a call to arms, Osler noted that his textbook of medicine was based on “personal experience correlated with the general experience of others.”1 This is a far cry from practice based on randomized clinical trials or even scientific evidence from observational studies. With all due respect to Osler, it is a fact that his method of practice and learning should have little place in contemporary medicine.

Patients would find it almost laughable that the evolution of current medical care has largely followed a “trial and error” pattern, driven by an educational system that was centered on the apprenticeship model. Our forefathers were educated based on the clinical experience of their mentors, and propagated the perceptions, attitudes, and behaviors with which they were presented. While exceptions to this model clearly existed, it has only been in the very recent past that the medical literature began to demand careful statistical analysis of published conclusions. The era of rigorous review of published material marked the beginning of a sea change in medical thinking.

Information in medicine creates advantage, whether that is due to more rapid diagnosis and treatment or clarity of understanding that opens the doors to discovery and progress. The trouble has been that the discipline of medicine is so broad that no single physician could master every facet and be expected to constantly integrate the rapidly changing landscape of new findings. The effective practice of medicine requires that new information be constantly incorporated, and yet the information sources that clinicians typically use fail to provide enough valuable data. To correct this problem, the notion of “evidence-based medicine (EBM)” has grown since its inception in the early 1990s. Evidence-based medicine was so-named as a means of communicating high-quality information to busy clinicians who were overwhelmed by the burgeoning body of literature in their field.

One widely published definition of EBM is “the conscientious, explicit and judicious use of the current best evidence in making decisions about the care of individual patients.”2 This definition highlights a critical piece of the EBM paradigm that is easily overlooked: medical evidence is generated based on evaluations of multiple measurements and yet, this information must be applied in the singular. Phrased another way, if we tell a patient that they have a 35% chance that their cancer will return after treatment, is it unreasonable for them to reply, “Which 35% of my body will the cancer involve?”

Evidence-Based Medicine [is] the process of systematically finding, appraising, and using contemporaneous research findings as the basis for clinical decisions.  Evidence-based medicine asks (sic) questions, finds and appraises the relevant data, and harnesses that information for everyday clinical practice.  Evidence-based medicine follows four steps: formulate a clear clinical question from a patient's problem; search the literature for relevant clinical articles; evaluate (critically appraise) the evidence for its validity and usefulness; implement useful findings in clinical practice.3

Public perception of medical care has evolved over the convoluted history of medicine. Surgeons no longer are double as barbers, and the general distrust of inpatient medicine has been replaced with an apparent reliance on the supremacy of the hospital. The modern era of medicine has heralded a widespread respect and wonder at the seemingly endless series of “miracles” that characterize progress from the days of Lister and childbed fever. In the current market of aggressive competition for faculty and patients, the role of information in medicine has never been more pronounced.

Radical changes are in motion in health-care delivery. So far they have been initiated externally and are often based on business and cost considerations. Much of this has been out of the individual physician's control to say nothing about their understanding. As a result, health care has evolved from a cottage to a mainstream industry and the physician has largely been left out of this evolutionary loop. Is there some fault on the physician's part that has allowed this loss of control? If so, what is it? Is it possible that maximum physician independence, revered above all else and at all costs, has undermined the ability for physicians to regain control of health-care delivery? If physician independence means preserving unexplained practice variance, which is currently out of control, perhaps physicians are their own worst enemy.

W. Edwards Deming, a statistician and management leader who taught quality as a system to the Japanese, spoke of variance management as being at the very core of quality improvement.4 Of course, in contrast to manufacturing, zero tolerance for variance and defects is not often possible in medicine. The focus must instead be on narrowing the range between upper and lower limits of a process. To this end, when a procedure or diagnostic test falls within a reasonable range of indications and evidence, the system process is in control. When procedures performed fall well outside this range, or if the range is exceedingly wide, the system process is out of control. Such is the case with health care, both on a micro- and macroscale. For decades, many studies have documented overuse of specific medical services, which are sometimes solely based on geographic location (e.g., certain cities have higher coronary bypass surgery rates with no local increased epidemiologic risk factors to explain this). Essentially, these practice patterns are “unexplained” by reasonable common causes or indications. In other words, reasonable variation in practice patterns is exceeded too often. One study specifically noted that 10%–27% of hysterectomies among women enrolled in seven health plans were performed for totally inappropriate reasons.5 Multiple studies performed by the RAND Corporation and others document similar findings and illuminate a quality problem in that patients are being subjected to risk of adverse consequences without documented benefit.6 It is worth repeating that the goal is not to achieve zero variance, rather merely to reduce practice variance to levels supported by existing evidence of benefit.

Perhaps we are asking the wrong primary questions and setting the wrong goals. For example, a commonly debated question is, “how can we reduce costs?” Should we instead be asking how we could improve the quality of care, minimizing unjustifiable and unexplainable practice variance? Is it possible that costs could be effectively reduced as a byproduct to this alternative and more palatable primary question? If the correct goals were set, could there be an alignment of effort toward true physician driven optimally managed care?

Is practice variance reduction and quality improvement via evidence-based guidelines just another medical “management fad” being foisted upon physicians? After all, there is a veritable alphabet soup proliferation of managed care and business based “buzz” words, reflecting the driving forces in the evolution of health-care delivery, usually not in concert with medical terminology: CQI, TQM, PDCA, EBM, LOS, Juran, Deming, etc. There are also more than 1500 practice guidelines in print and countless critical pathways. Most organizations seemingly have generated a set, but practical impact has been underwhelming. Can these help? How exactly? Are guidelines the same as critical pathways? Are these pathways just a “nursing thing”? How can we expect externally applied forces, embodied by practice guidelines forged by “expert” consensus panels, to be truly incorporated into better practice patterns at the local level? Certainly, unless the very best concepts in guideline development and evidence-based practice are actually implemented, there would be no net effect in patient care and cost efficiencies. Thus far, in fact, there has been no net impact. Perhaps we are not only asking the wrong questions but also taking the wrong approach. Perhaps the best solutions will come from individual physicians self-adjusting practice patterns based on evidence and outcomes rather than directives mandated from externally generated guidelines.

Is there a paradigm among all of the confusion that is workable? Is there a means by which the positive thrust toward value-added (outcomes/cost) care, as opposed to cheapest care, could be directed by physicians? One answer might be to practice the “best medicine” possible. Obviously, this is not a new concept. However, as we generate more and more data and publish it in a forest of journals, the ability to keep up gives way to information overload. Additionally, not all data has the same strength and the quality varies, and thus the data may not translate into good practical information. Along these lines, the randomized controlled clinical trial has been a gold standard and usually carries more weight than a case report/series or a consensus panel report. However, despite the estimated 250,000 or so randomized clinical trials that have confirmed the relative efficacy of many treatments, there is still a paucity of such studies to guide certain medical practice decisions. Without actually reading every single published journal, often in specialties that do not normally cross our desk, it is simply impossible for the average clinician to keep up with what is proven and what is not, and with what strength of evidence.

Ideally, to optimally obtain and use information, physicians would need electronic databases with continually updated data that is properly analyzed and processed. Most, if not all, practicing physicians either don't have this readily available or are not facile in searching current literature.7 Problem-based electronic database searches, using tools such as the MEDLINE, to answer specific questions are still infrequently used and may be incomplete since full article text may not be immediately available. In practice, the tendency is to evaluate problems based only on personal and often antiquated experience, or refer to respected authorities.7 Commonly, we refer to readily available resources such as textbooks to answer specific questions. The challenge inherent to this approach is in maintaining a full and up-to-date library, which is usually impossible.

Printed text, such as this book, will be criticized as being outdated no sooner than its publication. This is a potentially valid criticism. Much of what is summarized in this textbook is in fact not new, and in some areas quite dated but axiomatic. Nevertheless, it represents the best evidence known to the contributors to date. As long as one is aware of this limitation, by the 80/20 rule, it is generally true that 80% of evidence is relatively static and 20% represents new findings. In most cases, it is really the former scenario that is the bigger problem in practice variance minimization. In many instances, the main point is the lack of data to support efficacy of generally accepted common interventions. Of greater concern, despite strong data to support one point or another in patient care improvement, is that many physicians continue to practice status quo simply because “that is the way we have always done it.” In some cases, the available evidence is rapidly evolving and in other areas good solid evidence in existence for years or even decades has not been incorporated into general practice.

Are we discussing “cookbook medicine” here? Not at all. Medicine is still both an art and science and will continue as such so long as we treat human beings and not machines or biomechanical hybrids. However, we now have the operational tools to maximize the science while still supporting the art of delivering compassionate and effective care. Some of these tools are introduced in this text and the best evidence for perioperative and supportive care issues is presented. Selected chapters contain more axiomatic material than others. We have attempted to highlight controversial areas. However, this is a synopsis of evidence regarding general principles of perioperative care. As such, this textbook is not intended to be comprehensive and the reader is referred to the multiple excellent references within each chapter or to other works.

For all of the major technological advances in medicine, it is the struggle between art and science that defines medicine and separates it from aviation or manufacturing “widgets.” EBM demands that clinicians integrate the best available information on the behalf of their patients, simultaneously considering the individual patient factors and the best available evidence pertinent to the clinical situation. It is the critical thinking of the physician that allows EBM to thrive and acts as a counterbalance against practice limited by the most proximate clinical experience. Good physicians can integrate individual clinical expertise and the best available evidence because neither alone is enough. The centrality of this concept is summarized expertly by Sackett2:

Without clinical expertise, a practice risks becoming tyrannised by evidence, for even excellent external evidence may be inapplicable to or inappropriate for an individual patient. Without current best evidence, practice risks becoming rapidly out of date, to the detriment of patients.

PERIOPERATIVE AND CRITICAL CARE ECONOMICS

At the societal decision-making level, we must ultimately balance the focus on maximum acute care with optimal care of an aging population. While some chronic and catastrophic diseases, such as cancer, often cannot be cured, patients still need relief from symptoms and minimization of disease-related complications and dysfunction. All of these issues touch upon supply and demand realities and opportunity costs. Maximizing efficiencies and minimizing unexplained practice variance will go a long way toward conservation of scarce resources and improved outcomes, and will contribute to overall cost reduction in health-care delivery services.

If one adopts a classic economic marginal analysis approach toward mortality, morbidity, level and extent of ICU care, complications, and avoidance thereof, one can isolate the incremental impact of each decision on quality and costs. Key questions might be as follows: During preoperative evaluation, prevention of complications is key to decreased morbidity and length of stay. How much more, or perhaps less, is required as a diagnostic input to achieve a given superior quality output? During perioperative care, what incremental opportunities exist for prevention with appropriate surveillance and management that are based on good evidence? In using new technologies, or even older technologies, what is the appropriate incremental use of such resources and what are their limitations toward optimizing outcomes?

EBM exists then as a tool, but not a substitute for the clinician. The proper implementation of EBM has driven the growth of an entire field of medicine, replete with its own experts, critics, proponents, and detractors. The most widely understood conceptual framework of EBM is the idea of the hierarchy of evidence. This is referred to as the first fundamental principle of EBM, and is discussed in more detail in Chapter 2. Even though randomized clinical trials are the gold standard, there absolutely is a place for cohort studies, case–control studies, and other “lesser” evidence. This does not mean that issues that are not defined by randomized trials are incompletely tested. All levels of evidence are worthwhile, and best evidence is exactly that—the best that can be generated for a topic. The second fundamental principle of EBM is the idea that regardless of the level of evidence, value and preference judgments are implicit in every clinical decision.8

It is important to recognize that EBM is not without its limitations. As with any effective tool, it is perhaps easier to misuse rather than properly utilize it. Tragedy can result from paying attention to poor quality evidence instead of good quality evidence, and critical appraisal cannot be abandoned for blind acceptance. Many medical schools and training programs, in a form of premature closure, are moving away from teaching the fundamentals of careful evidence appraisal to emphasize the implementation of evidence. The intent of this new focus is to produce high-quality, safe, and low-cost care (i.e., Accreditation Council for Graduate Medical Education competencies of systems-based practice and improvement and practice-based learning). However, abandoning appropriate skepticism regarding the effectiveness of these interventions may lead to large investments in quality improvement, safety, and efficiency activities that fail to yield the expected benefits.

The same can be said for incorrectly applying population-based models to best practices in individual care. Major pronouncements about a particular action or intervention are not served by EBM, and those who try to misuse the literature in this way risk harming the very group that they have sworn to protect. Regardless of one's perspective on EBM, the discipline reflects the desire of all involved in patient care to improve the quality of patient care.

OPERATIONAL TOOLS OVERVIEW

This textbook strives to present the best available evidence for decision making in perioperative and supportive care in the gynecologic oncology patient. It also introduces some operational mindsets and tools. Questions that this textbooks addresses include the following:

What is evidence-based medicine? How does one find all the available data? What if there is no good data? How does one evaluate which evidence is best for the given situation?What evidence exists toward minimizing unexplained variance and optimizing practice patterns?Are there any formal decision analysis methodologies that can help?Can these principles translate into practice guidelines that can actually be implemented and contribute to improvement?Is this just cookbook medicine? Or, is it a guide toward evolution of best practices specific to each physician's and patient's environment?

This textbook primarily addresses gynecologic oncology care, but can readily apply to complicated gynecologic perioperative care. Key issues are associated with a level of evidence score within the text or in algorithm form. Some chapters also contain more axiomatic information than others, and as such are not always subject to grading. In other subject areas the lack of extensive underlying evidence is striking. Chapter 2 addresses information gathering and interpretation tools. The subsequent clinical chapters present the contributing authors’ best efforts to gather and synthesize up-to-date information addressing best approaches to common as well as uncommon problems in perioperative, supportive, and critical care. Some authors found data gathering and grading more second nature than others and so some biases remain. These areas should be apparent and interpreted to mean that the subject area is heavily influenced by level III data. Editing cannot always alleviate this and may confuse the reader if expert opinion meaning is altered.

This second edition is constructed largely as the first. Many of the chapter topics have remained the same, and others have been updated to include major changes in practice since the publication of the first edition. Some chapters have been deleted to narrow the focus on the perioperative nature of the text, and new chapters have been included in herbal and complimentary medicine, end-of-life decision making, and fertility-specific issues pertinent to the gynecologic oncology patient.

In summary, this is an imperfect but focused and genuine effort to present the best available information designed to help decrease practice variance toward predictable improved clinical outcomes. As presented, it is anticipated to be a kernel work in progress, constantly improved through revision, and a guide for reader-directed updating and local adaptation.

REFERENCES

1. Osler W. The Principles and Practice of Medicine. 8th ed. New York: Appleton and Co.; 1918.

2. Sackett DL, Rosenberg WM, Gray JA, et al. Evidence based medicine: what it is and what it isn’t. BMJ 1996;312(7023):71–72.

3. Rosenberg W, Donald A. Evidence based medicine: an approach to clinical problem solving. BMJ 1995;310(6987):1122–1126.

4. Walton M. The Deming Management Method. 1st ed. New York: Putnam Publishing Group; 1986.

5. Bernstein SJ. The appropriateness of hysterectomy: a comparison of care in seven health plans. JAMA 1993;269:2398–2402.

6. Chassin MR. Assessing strategies for quality improvement. Health Affairs 1997;16:151–161.

7. Olatunbosun OA, Edouard L, Pierson RA. Physician's attitudes toward evidence based obstetric practice: a questionnaire survey. BMJ 1998;316:365–366.

8. Guyatt GH, Haynes RB, Jaeschke RZ, et al.; for Evidence-Based Medicine Working Group. Users’ guides to the medical literature: XXV. Evidence-based medicine: principles for applying the Users’ Guides to patient care: JAMA. 2000;284(10):1290–1296.

Chapter 2

Evidence-Based Medicine and Decision Support

Steven A. Vasilev, MD, MBA

WHAT IS EVIDENCE-BASED MEDICINE?

Bertrand Russell noted, “The extent to which beliefs are based on evidence is very much less than believers suppose.” It is easy to see how this may be directly applicable to medical practice.

Under the protective blanket of the “art” of medical practice, decision making has often been based on anecdotal experience and incomplete utilization of the best available objective data. Physicians tend to practice based more on what their attendings taught them rather than continually researching existing evidence.1,2 It then becomes a matter of unchanging practice routine to make decisions based mostly on personal experience and intuition. In effect, every physician may be doing what they perceive to be their best, but that is not enough. Data suggest that many decisions are made contrary to available evidence and up to half of decisions are based on weak or no evidence.2–4 In light of this, it is essential that the medical educational process be carefully scrutinized to see if it requires transformation to optimize decision making and reproducible quality outcomes.5–7 Largely due to inattention to the above, unexplained practice variance is a significant problem, contributing to runaway costs and a wide range of outcomes.7–14

W. Edwards Deming, a leader in cross-industry quality issues, stated that all of his work at the core was based on controlling variance via statistical process control (SPC).15 What is unexplained or assignable clinical practice variance? In any industry, to maintain the quality of outcomes, operations and processes must be continually inspected and tested to minimize defects. In health care, just as in other industries, there will be variability in outcomes. The goal in health-care services might not be zero defects as proposed in some manufacturing scenarios, but rather minimization of variability in outcomes such that the result as a whole is of acceptable quality. The question then becomes “Is the outcome variability following a given intervention due to chance (random) variation or assignable (nonrandom) unexplained variation?” Chance variation is that which is built into a system, such as the range of normal hemoglobins in the physiologic system of a patient. Assignable variation occurs if some portion of the system is out of control and is amenable to intervention. One of the biggest challenges is to determine if a process or activity is out of control, requiring adjustment, or not.

During the 1920s and 1930s, while at Bell Telephone Laboratories, Walter Shewhart developed statistical control charts to help determine when assignable variation has occurred. A repetitive operation, such as caring for a set of patients with a particular problem, will seldom, if ever, produce exactly the same result. However, the outcome variability surrounding a mean value and standard deviation will often produce a normal distribution for the population. As in any other statistical scenario, periodically examining the entire population of interest for variation in the mean is not feasible. Instead, sampling is performed along with selection of an upper control limit (UCL) and a lower control limit (LCL). If the sample mean exceeds control limits, typically set at +/–3 standard deviations in most industries, the possibility that the variation is due to chance is <0.3%. Taken one step further, it is even possible to predict assignable variation prior to loss of system control. Also, in some cases, assignable variation can represent improvement in process rather than loss of control and should thus be investigated.

The above brief discussion is meant to introduce the idea that unexplained clinical practice variance should be sought, the reasons identified, and a process improvement implemented. This could be on an individual practice or organizational or even societal level. Details regarding statistical process control are beyond the scope of this textbook, and the interested reader may refer other sources.16 The key point is that tools are available to assess how unexplained variance can influence outcomes and point to areas requiring attention and correction.

Unexplained variance due to assignable variation has the potential to be significantly influenced by evidence-based medical practice. How scientific can the base for the art of medicine get? During the 1990s, it was estimated that less than 10% of clinical practice was based on solid randomized controlled clinical trial (RCT) data, and more recent reports suggest that many guidelines are still of dubious quality.3,4,17–19 While this may be a criticism of evidence-based medicine (EBM), it reflects a core misunderstanding of EBM intent.20,21 In fact, the other 90% of clinical practice should still be based upon the best available level of evidence or at least the understanding that a particular practice pattern or intervention is not well grounded in evidence of any kind.18,22 This understanding alerts clinicians that their convictions may be on shaky ground and that further evaluation may be required to minimize practice variance.

What is more disturbing is that the 10% of clinical practice that is well grounded in solid RCT data is often not incorporated into standard patient care, and unexplained variance is seen in this group of interventions as well. Lag time between publication of compelling data and incorporation can exceed 10 years.23

Is EBM a new “managed care” concept? On the contrary, the concept of EBM dates back more than a century.24,25 Recently, largely due to “managed care” pressures, it has enjoyed a resurgence of interest among divergent groups: clinicians wishing to regain control of medical practice, payors wishing to limit variance and lower costs, health-care purchasers wishing quality initiatives for better value, and the public wishing to understand what the “best” approach really is.26 Numerous centers and Internet Web sites for evidence-based practice have been established in numerous disciplines. For example, the Cochrane Collaboration, an international multicenter venture with an internet portal (www.cochrane.org), reviews, synthesizes, and distributes data on health-care practices on an ongoing basis.27,28 A number of EBM journals have also been introduced. Additionally, recommendations for improvement in reporting of clinical trials have been proposed and endorsed by leading journals.29,30 These recommendations are embodied in CONSORT or Consolidated Standards of Reporting Trials (www.consort-statement.org). Related forms of standardization are proposed, such as STROBE or Strengthening the Reporting of Observational studies in Epidemiology (www.strobe-statement.org), QUOROM or QUality of Reporting of Meta-analyses, and STARD for diagnostic studies (www.stard-statement.org).31

This resurgence of interest in EBM has certainly faced significant criticism.32–35 Some have interpreted it to be a fad, a pure cost cutting device, while others label it “cookbook” medicine, which at its core is an antithesis to the “art” of medicine. Much of this criticism and negative reaction is based upon the misunderstanding of terminology and philosophy. It may also be negatively associated with the decade-long continued backlash against “managed care.” However, lost among this is the EBM philosophic thrust toward improvement in the quality of medical practice. Although some believe that quality de facto requires higher costs, the converse is likely true in most cases. There is no question that technology transfer will continue to affect health-care costs. However, the correct application of new and emerging technology and limitation of poor management decisions will serve as a counterbalance. In the optimal scenario, overall quality improvement will decrease costs.36,37 Additionally, convergence of information technology and computer-assisted decision analysis and support will likely accelerate and contribute to quality improvement and cost control.38 Thus, it may be axiomatic that true improvements in quality and efficiency will control costs. As long as quality is better defined and becomes the lead issue, and costs are well described and assigned, a relationship between the two will become easier to evaluate and implement process improvement. Currently, neither is the case. Quality is not clearly defined and costs are usually improperly assigned.39

Thus, EBM may be poised to utilize the boom in information technology and improve quality through unexplained variance reduction, but it is certainly in a state of evolution. Various interpretations and local adaptations are apparent and will continue while health care is in a state of reorganization. With the above in mind, it may be helpful and illuminating to review what EBM can do and what it will not do.

At its core, EBM may be viewed as a process. This process combines systematically obtaining the best available refined data (i.e., information) on a defined topic with the clinical expertise of the clinician. The resulting knowledge base then is applied to patient care.5,40 While this may seem like good old-fashioned medical care, the data gathering portion of the process is the first significant challenge. In today's environment of data overload, physicians are challenged to keep up with the myriad of peer-reviewed journal articles, much less assess the adequacy of each study and the strength of evidence presented. A new paradigm and skill set is required, including efficient access to published studies and systematic application of evidence strength analysis.41

The second challenge is to continually and systematically incorporate the best evidence into clinical practice. The application of clinical expertise and judgment, heretofore known as the “art” of medicine, is an integral part of compassionate health-care delivery and cannot be summarily replaced by data driven guidelines. Physicians must continue to combine research-based evidence with accumulated clinical expertise.42–44 The physician is a knowledge worker and the goal should be the development of a learning system within the individual a la a Peter Senge modification of the learning organization, which is based upon continually improving systems.45,46 In a patient-centered environment, hard objective data may point toward one intervention but, based on clinical experience and patient input, the best outcomes may be realized by taking a different path for a given clinical situation. Optimally, both data and expertise should be available, interventions applied and outcomes assessed.

A common misperception is that EBM is equivalent to cookbook medicine. Pat prescriptions for any imaginable condition without regard for individual patient requirements or physician's clinical expertise would fit that definition. However, EBM mandates integration of clinical experience and a commitment to using the best available data as a guide to continued improvement and incremental narrowing of practice variance patterns.

EBM cannot and does not have cost reduction as its primary goal. Rather, more effective utilization of diagnostic and treatment resources has as its by-product possible cost reduction. This does not always occur and depends upon how costs are defined and to whom they are assigned.

EBM does not rest entirely upon external research findings with the RCT as the gold standard. While it is true that well-designed and interpreted prospective studies should have a strong influence on shaping practice patterns, they are not the end all. Neither is the compilation of the same via meta-analyses. The process goal is to define the best available external evidence that addresses specific clinical questions. The type of evidence depends on the clinical question at hand. Some questions require a RCT, while others require a cohort study, and still others may merely require an observation of widely disparate efficacy of a given intervention.

Traditional CME has failed to keep physicians abreast of new developments to the extent that these innovations and developments are not readily integrated into practice.47–49 Worse, the half-life of medical education is getting ever shorter. Without the best current evidence, practice patterns can become dangerous and costly in both an economic and morbidity sense. EBM can help maintain the balance between data-driven guidelines and patient centered “art” of medical practice.50

EBM is a philosophic approach to medical practice and education. It is not outcomes research, but can help delineate where evidence is lacking as a basis for future research. It does so by delineating what kind and strength of evidence exists for a particular question and facilitates the gathering and grading of such evidence. As more supporting evidence becomes available, the validation of EBM as a philosophy and generally accepted practice infrastructure appears to be warranted.51,52

So how does EBM get incorporated into everyday medical practice? According to Sackett, there are five key steps to incorporating evidence-based practice into day-to-day medical care. First, the clinical problem must be framed into answerable questions. Second, the best available data, which will help answer the question, must be quickly and efficiently tracked down. Third, the evidence must be critically appraised for validity and usefulness as information. Fourth, the appraisal results must be incorporated into daily practice. Fifth, the results of this change in practice patterns must be evaluated for outcomes and other feedback regarding the new practice pattern.53

EXTERNAL DATA GATHERING AND RELEVANCE INTERPRETATION TOOLS

As the information age explodes, the total amount of readily available knowledge far exceeds the clinical experience of a single physician or consensus group of experts. Additionally, due to the electronic superhighway and multiple electronic storage media, one no longer needs to read through masses of journals and memorize key references in order to stay current.54,55

The following represents a synopsis of helpful sources and strategies for searching the medical literature. For a more comprehensive discussion, refer to the classic Sackett's Evidence Based Medicine: How to Practice and Teach EBM.53

In general, evidence must be current and credible, applicable to the practitioner's patient population and clinically relevant above and beyond an acceptable p value or confidence interval.

Electronic searches have practically become synonymous with MEDLINE or PubMed searches. As described below, MEDLINE is the largest and most well-known database. However, other sources may be better for focused topic searches. For example, recent research topics and cutting edge information may be best found in Current Contents. On the other hand, for a general review on a particular topic, possibly with CME credit attached to help fulfill licensing requirements, Medscape or other online journal databases should be consulted (Table 2.1). Online structured reviews and guidelines provide an evidence-based synthesis of available literature on a growing number of topics (Table 2.2). The overview given in this chapter is pragmatically structured around latest and most useful Web sites rather than just providing artificial constructs of database vs. search engine classifications

Table 2.1 Evidence-Based E-Journals

Core Evidencehttp://www.coremedicalpublishing.comAnesthesia and Intensive Carehttp://www.aaic.net.au/home.htmlAnesthesiologyhttp://www.anesthesiology.orgACP Journal Clubhttp://www.acpjc.orgPIERhttp://pier.acponline.org/index.htmlJAMAhttp://www.ama-assn.org/public/journals/jama/jamahome.htmNew England Journal of Medicinehttp://www.nejm.org/EBM Onlinehttp://ebm.bmj.comCochrane Collaborationhttp://www.cochrane.org/index.htmlUpToDatehttp://www.uptodate.com

Table 2.2 Online Structured Reviews and Guidelines

Society of Critical Care Medicinehttp://www.learnicu.org/Quick_Links/Pages/default.aspxEvidence-Based Medicinehttp://www.acponline.org/clinical_information/guidelines/current/#acgJAMA Evidencehttp://www.jamaevidence.comHealth Services Technology Assessment Text (HSTAT)http://text.nlm.nih.gov/Cochrane Collaborative databasehttp://www.cochrane.org/index.htm

MEDLINE

The most widely available tool to clinicians, and also available directly to the public, is the searchable MEDLINE database through the National Library of Medicine. Most electronic search engines that access this online portion of the National Library of Medicine retrieve references by textwords of medical subject headings (MeSH). Searches can be initiated for specific article titles, authors, general subjects, or specific clinical questions.

The most popular and easy to use search engines are PubMed and Ovid via Universities. Abstract availability is free through the NLM (http://ncbi.nlm.nih. gov/PubMed). Abstracts are available for over 75% of the references, comprising 19 million citations from MEDLINE and life science journals indexed from 1966 on.

Basic text keyword searches can certainly provide reams of basic and broad information. Unfortunately, at this point there is so much information published that there is pollution and overload. It is becoming difficult to tell the difference between good and misleading or low quality information. To help address this, search strategies can significantly influence the quantity and quality of data found.56 First, the MEDLINE database should reflect the available literature for the appropriate time period.57 Second, all evidence for a given topic should be considered, from case series or consensus opinions up to randomized double blind prospective studies. Third, all evidence must describe the specific patient population or clinical problem in reasonable detail. This centers upon entry of the appropriate key words and MeSH headings. Since the early 1990s the MeSH vocabulary has expanded to include specific search string toggles for type of study (e.g., controlled trial vs. cohort) and publication (e.g., meta-analysis).58 Although validity of studies should be individually appraised, by including these methodology toggles in their search, users of the MEDLINE and other online NLM databases can search by type of studies available. The Health Information Research Unit (HIRU) of McMaster University (http://hiru.mcmaster.ca/hiru/) has suggested strategies for optimal retrieval of relevant citations by methodology based MEDLINE searching.56 For example, in order to maximize the proportion between relevant and irrelevant citations retrieved, the search strategy (included with other key words or MeSH headings) EXPLODE SENSITIVITY and SPECIFICITY OR PREDICTIVE VALUE (TEXTWORD) can be used. If a personally performed search does not provide appropriate information, librarian consultation should be sought.

Online Journals and Databases

A very brief and incomplete list of available journals most closely pertaining to evidence-based perioperative care is given in Table 2.1. Most of these journals come with an online search engine, many are free or have featured evidence-based article reviews and indexed back to 1995 or earlier.

Textbooks

Textbooks have the distinct advantage of presenting synthesized information on various topics and the information is quickly accessible. The disadvantage lies in the danger of being outdated as of the publication date. Unfortunately, as of a 2005 review of medical library researchers found that physicians still rely primarily on two information resources: colleagues and printed textbooks and journals.59 This practice has not changed much since 1992, when a similar review was conducted. However, if EBM philosophy is applied and explicit links to evidence are identified, textbooks are still a valuable resource. In particular, there is relative stability of supporting data for 80% of clinical interventions. The other 20% is subject to periodic change, although this ratio may change depending upon how dynamic the given area is. Thus, a textbook synthesis of EBM information can facilitate incorporation into clinical practice those interventions that are clearly well grounded and facilitate removal from clinical practice those interventions that are clearly without basis. In order to capture a current compendium regarding clinical problems and interventions, rapidly updated CD-ROM or online textbooks offer a distinct advantage.

Meta-analysis

A meta-analysis is a secondary systematic review of published primary data, optimally including RCT. Often smaller studies are grouped together in order to apply statistical analysis tools to an increased pooled sample size.60 While meta-analysis has gained widespread acceptance, the utility of this method has been questioned.61,62 Standardized approaches have therefore been recommended.31 Additionally, the results are limited by the quality of the primary data publications. We know that large randomized trials and systematic reviews disagree at a rate greater than chance alone would predict. Therefore, the review must include a description of how the primary data sources were identified, and provide enough information for the reader to be able to assess the possibility of selection and analysis distortion due to various biases. Continuous compilation of clinical trial results and analysis of the results have been suggested and are provided on multiple subject areas by the Cochrane Collaboration.63 Although not meta-analysis per se, the ACP Journal Club provides online abstracted reviews of primary literature.

EBM Guidelines

Evidence-based guidelines represent an attempt to distill the best available evidence via a structured process for a particular clinical question.64 Many, if not most, published guidelines, which number in the thousands, do not employ such a rigorous process and should be interpreted accordingly.65 Guidelines written by external sources also do not enjoy widespread implementation. The GRADE System (Grades of Recommendations Assessment, Development and Evaluation) was introduced in 2004 and endorsed by many organizations, including the Cochrane Collaboration. There is concern that this methodology for rating guidelines is yet to be validated.66 However, with these caveats in mind, the Society of Critical Care Medicine operates a good online guidelines site pertaining to issues in perioperative and critical care. Additional resources are listed in Table 2.2, including HSTAT (which provides NIH Consensus Statements and AHCPR Evidence Based Guidelines) representing the two extremes of guideline development process.

Grading Published Data

Once a representative data set of published references is obtained, it must be systematically reviewed to determine the quality of data presented. The key question is: “which data set translates into the best information available for the specific question at hand?” Best may be defined as relevant and with the least methodological flaws. These flaws differ depending upon the type of study design, and are introduced below. Extensive discussion regarding clinical research design is well beyond the scope of this textbook, and there are many excellent textbooks and reviews67–71 to refer. A particularly comprehensive review series of articles, entitled “User's Guide to the Medical Literature”, published by the Evidence-Based Medicine Working Group from McMaster University72–85 is highly recommended.

The individual studies must then be evaluated for strength of evidence and categorized by evidence-based medicine (EBM) level. This textbook refers to guidelines proposed by the U.S. Preventive Services Task Force (USPSTF).86,87 Although this system was proposed for the evaluation of clinical preventive services, it is generally applicable to other interventions. These EBM levels may be noted in the text or in the decision algorithm or both, depending upon the chapter and material (e.g., level I; 3 and Level II-3; 6). When available, the number of “best” studies/reports used to support the evidence are noted after the semicolon.

Level I: Evidence obtained from at least one properly designed and powered RCT; well-conducted systemic review or meta-analysis of homogeneous RCTs.Level II-1: Evidence obtained from well-designed controlled trials without randomization.Level II-2: Evidence obtained from well-designed cohort or case–control analytic studies, preferably from more than one center or research group.Level II-3: Evidence obtained from multiple time series with or without the intervention. Dramatic results in uncontrolled experiments could also be regarded as this type of evidence.Level III: Opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees.

Other grading systems also exist, such as the Canadian Task Force (www. ctfphc.org) and the Centre for Evidence Based Medicine (CEBM) at Oxford (www. cebm.net), but are similar in nature.

A letter grade may then be assigned, summarizing the overall strength of evidence. Most grades are specific for the preventive services or screening question at hand. Nonetheless, a general appreciation for the use of levels of evidence as they translate into “grades” is reviewed below. In this textbook, letter grades are not assigned, leaving the reader to make their own assessment based on EBM Level of evidence and applicability to their patient population. The USPSTF made some changes to its grade definitions and introduced a “statement of net benefit” in May 2007. This should be borne in mind when reviewing grades assigned before and after this date. The following represent grades using information from both USPSTF and the CEBM systems.

Grade A: Consistent level I studies (CEBM). The USPSTF recommends this service. There is high certainty that the net benefit is substantial.Grade B: Consistent level II or III studies or extrapolations from level I studies (CEBM). The USPSTF recommends this service. There is high certainty that the net benefit is moderate or there is moderate certainty that the net benefit is moderate to substantial.Grade C: Level IV studies or extrapolations from level II and III studies (CEBM). The USPSTF recommends against routinely providing this service to an individual patient. There is at least moderate certainty that the net benefit is small.Grade D: Level V studies or troublingly inconsistent or inconclusive studies at any level (CEBM). The USPSTF recommends against the service. There is moderate or high certainty that the service has no net benefit or that the harms outweigh the benefits.I Statement (USPSTF only): The USPSTF concludes that the current evidence is insufficient to assess the balance of benefits and harms of the service. Evidence is lacking, is of poor quality, or is conflicting, and the balance of benefits and harms cannot be determined.

WHAT IS THE “BEST” EVIDENCE?

There is no consensus about which evidence rating system is ideal or which should be universally applied. Additional systems have been proposed, based on the field of medicine or are journal specific. You may find any of these, or a combination hybrid, in use by any given journal.

The “best” source of high quality evidence is generally considered to be the RCT. This study design provides very useful information with the least bias vulnerability, but it is not the appropriate answer to all questions. Many interventions or other clinical questions have not been, nor ever will be, investigated at that level. Or, the clinical question being asked may not be appropriate for evaluation in a randomized controlled fashion. Examples are abundant and include low prevalence conditions, treatment or diagnostic decisions of low impact, low morbidity conditions, and highly complicated very high cost low volume interventions. For example, the Pap smear was never evaluated by RCT. However, enough observational epidemiologic evidence exists and so its role in reducing the incidence of invasive cervical cancer is not generally questioned. Similarly, certain other questions require answers not from RCTs but from observational cohort and case–control studies, outcomes data within a local practice setting, questionnaire or interview-based data, problem modeling data, and other formats.

Intuitively, it is important to consider what type of information an RCT provides. Usually, the RCT compares two or more treatments, which are felt to differ relatively minimally in efficacy, in order to determine which is the better intervention. The true effectiveness of an RCT treatment arm is often inversely proportional to the number of RCTs it has been a part of. Thus, if a significant advance occurs, such as the introduction of the Pap smear or of antibiotics/penicillin, an RCT is not mandated to demonstrate efficacy.

EBM grading schemes do place high value on RCTs, but this should not overshadow the fact that some questions are better answered by alternative study designs. Total homage to and reverence for the “p value” is not realistic. Instead, asking the appropriate question usually determines the appropriate study design or data source and is a philosophic underpinning of EBM practice.

EVALUATING PROSPECTIVE RCTS (EBM LEVEL I)

Clinical trials are intervention based on definition, with the subjects prospectively assigned to experimental and control arms. The subjects in both arms should be similar in all or most characteristics, allowing outcomes of a given intervention to be accurately assessed between the arms. While blinded studies are more difficult to design and implement, unblinded studies overestimate benefit by up to 17%.

Major advantages of this study design are as follows. (1) Bias elimination by random assignment of intervention. (2) Facilitates blinding of the investigators as well as participants whenever possible, further reducing bias. 3) Facilitates statistical probability analysis in determining strength of cause–effect relationship.

Study design flaws to watch for may include the following. (1) Invalid methods of randomization. (2) Subjects not matched by key variables. (3) Sample size too small to detect potentially important differences. (4) Compliance problems or loss to follow-up rate affects the outcomes analysis.

EVALUATING CONTROLLED PROSPECTIVE NONRANDOMIZED STUDIES (EBM LEVEL II-1)

Level II-1 studies are also interventional in nature. However, the study and control groups are not prospectively and randomly assigned, introducing biases. When using historical or otherwise unmatched control groups, the reported effects of the intervention being studied are often exaggerated.88,89

Study design flaws may include those listed under the Level I category. The main detraction from the highest reliability rating is the nonrandom assignment between the study and control groups. This could unevenly distribute known and unknown variables and factors that can be significant enough to sway the outcome analysis and interpretation.

EVALUATING OBSERVATIONAL STUDIES (EBM LEVELS II-2, II-3, AND III)

These studies represent the bulk of evidence available for the majority of diagnostic and treatment interventions. They include both prospective and retrospective study designs as delineated hereafter.

Level II-2

Case–Control Studies

The design of these studies is generally retrospective. First, the research question is identified and explicitly framed. Then the objective is to identify subjects with and without a disease or condition. A sample population is identified as well as a matched sample population without the condition but with similar demographics. The background of each group is then reviewed in order to find out why the cases developed a given condition whereas the controls did not. The odds that a given intervention or exposure produced the condition in the case group is compared with the control group. As such the output of the study is the odds ratio, which approximates the relative risk rather well if the prevalence of a condition is not excessively high. The major advantages are as follows. (1) The study can be rapidly performed. (2) It is useful when studying rare conditions.

Major study flaws to watch for include (1) unknown factors that influenced development of a condition inadvertently not taken into consideration, (2) selection bias preferentially included or excluded subjects in either the case or control group, and (3) differential recall bias may affect the ability of either group to remember, admit to, or include certain key exposures being investigated.

Cohort Studies

Cohort studies can be either prospective or retrospective. In each case, a group of subjects is followed through time, usually made up of exposed and unexposed subjects to a given risk factor or intervention. In a prospective