PA I N M E DICI N E B OA R D R E V I E W

PAIN MEDICINE BOAR D

R EV IEW

EDITED BY

Marc A. Huntoon, MD
PROFESSOR OF ANESTHESIOLOGY

V IRGINI A COM MON W EALTH UNI V ER SIT Y

RICHMOND VIRGINIA

1
 1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.

Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.

© Oxford University Press 2017

All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.

You must not circulate this work in any other form

and you must impose this same condition on any acquirer.

Library of Congress Cataloging-​in-​Publication Data

Names: Huntoon, Marc A., editor.
Title: Pain medicine board review / edited by Marc A. Huntoon.
Other titles: Pain medicine (Huntoon)
Description: Oxford ; New York : Oxford University Press, [2017] |
Includes bibliographical references and index.
Identifiers: LCCN 2016029256 (print) | LCCN 2016029915 (ebook) |
ISBN 9780190217518 (alk. paper) | ISBN 9780190217525 (e-book) |
ISBN 9780190217532 (e-book) | ISBN 9780190217549 (online)
Subjects: | MESH: Pain | Pain Management | Examination Questions
Classification: LCC RB127 (print) | LCC RB127 (ebook) | NLM WL 18.2 |
DDC 616/.0472076—dc23
LC record available at [Link]

This material is not intended to be, and should not be considered, a substitute for medical or other professional advice.
Treatment for the conditions described in this material is highly dependent on the individual circumstances. And, while
this material is designed to offer accurate information with respect to the subject matter covered and to be current as of
the time it was written, research and knowledge about medical and health issues is constantly evolving and dose schedules
for medications are being revised continually, with new side effects recognized and accounted for regularly. Readers must
therefore always check the product information and clinical procedures with the most up-to-date published product
information and data sheets provided by the manufacturers and the most recent codes of conduct and safety regulation.
The publisher and the authors make no representations or warranties to readers, express or implied, as to the accuracy or
completeness of this material. Without limiting the foregoing, the publisher and the authors make no representations
or warranties as to the accuracy or efficacy of the drug dosages mentioned in the material. The authors and the publisher
do not accept, and expressly disclaim, any responsibility for any liability, loss or risk that may be claimed or incurred as a
consequence of the use and/or application of any of the contents of this material.

1 3 5 7 9 8 6 4 2
Printed by WebCom, Inc., Canada
 DISCLOSUR ES

The views expressed in Chapters 12 and 24 are those of the authors and do not necessarily reflect the official policy or posi-
tion of the Department of the Navy, Department of Defense, or the United States Government.

v
 CONTENTS

Preface ix 14. Complementary and Alternative Medicine 190

Contributors xi Elizabeth Huntoon
15. Work Rehabilitation 202
1. Pain Anatomy and Physiology 1 Robert Yang
Daniel J. Pak and Neel Mehta 16. Acute Pain Management and Tissue Pain 215
2. Literature Review and Evidence 12 Ignacio Badiola
Andrea L. Nicol and Usman Latif 17. Cancer Pain 227
3. Pain Research: Placebo, Animal Models, Ethics, Amitabh Gulati and Joseph C. Hung
and Epidemiology 28
18. Cervical Radicular Pain 245
David A. Edwards
Joshua Horowitz
4. Psychology of Pain: Psychological Assessment
and Treatment of Pain 44 19. Low Back Pain and Radicular Pain 256
Julie R. Price, Micah J. Price, and Marc A. Huntoon Marc A. Huntoon
5. Gender Differences in Pain 63 20. Chronic Pelvic Pain 274
Gregory Carpenter and Meenal Patil Martha J. Smith
6. Imaging 70 21. Obstetric Pain 283
Markus A. Bendel, Drew M. Trainor, Hans P. Sviggum and Adam K. Jacob
and Susan M. Moeschler 22. Headache 300
7. Addiction and Pain 88 Kurt F. Dittrich
Daniel F. Lonergan 23. Orofacial Pain 315
8. Pharmacology 102 Christopher Sobey
Ryan Nobles 24. Neuropathic Pain 327
9. Miscellaneous Pharmacology 120 Ian M. Fowler, Robert J. Hackworth,
Ramana K. Naidu and Erik P. Voogd
10. Neuromodulation 131 25. Complex Regional Pain Syndrome 348
Bryan Covert and Marc A. Huntoon Jenna L. Walters
11. Pain Management Techniques 143 26. Pediatric Pain and Development of Pain Systems 357
Maureen F. McClenahan and William Beckman Ellen W. K. Rosenquist and Natalie Strickland
12. Musculoskeletal Pain 162 27. Geriatric Pain 374
M. Gabriel Hillegass, Anthony A. Tucker, Elizabeth Huntoon
and Antonio Quidgley-​Nevares 28. Pain Relief in Areas of Deprivation and Conflict 384
13. Physical Medicine and Rehabilitation and John Corey and Kelly McQueen
Electrodiagnosis 173
Aaron Jay Yang Index 395

vii
 PR EFACE

I undertook the writing of this book because of what I per- practice. The field of pain medicine has continued to evolve
ceived as a gap in the available board review books at the during the quarter century that I have been practicing,
time of my last recertification in pain medicine. There were a and the need for leadership in education has not lessened.
couple of books out then, but they seemed to be less focused Although no one book can be a sole source of study material
on the actual key words and examination content outlines for such an all-​encompassing specialty area, it is my hope
produced by the American Board of Anesthesiology (ABA) that this book will help medical students become interested
and other parent boards. As a former question writer for in the field and that residents, fellows, or recertification
the ABA exam, I was well aware of the goals of those who candidates would become familiar enough with the mate-
develop those board exams. Despite the dearth of avail- rial that they could pass the examination. I wish you the
able review books, I (fortunately) did not find the exam best, and hope that you find this field to be as fulfilling as
to be too difficult. However, as someone who has led pain I have, while remaining cognizant of the privilege it is to
medicine programs at institutions such as the Mayo Clinic, serve patients in pain. We must first be humble and strive to
Vanderbilt University, and now Virginia Commonwealth become knowledgeable to be the best we can.
University, one should expect that academic faculty would
stay abreast of the information relevant to a modern pain Marc A. Huntoon, MD

ix
 CONTR IBUTOR S

Ignacio Badiola, MD David A. Edwards, MD, PhD

Assistant Professor of Anesthesiology and Critical Care Assistant Professor of Anesthesiology
University of Pennsylvania Perelman School of Medicine Vanderbilt University
Philadelphia, PA Nashville, Tennessee

William A. Beckman, MD, CAPT, MC, USN Ian M. Fowler, MD

Assistant Professor of Anesthesiology Assistant Professor of Anesthesiology
Uniformed Services University of the Health Sciences Uniformed Services University of the Health Sciences
Staff Anesthesiologist and Pain Medicine Physician Staff Anesthesiologist and Pain Medicine Physician
Naval Medical Center Portsmouth, Virginia Naval Medical Center San Diego
San Diego, California

Markus A. Bendel, MD Amitabh Gulati, MD

Department of Anesthesiology Director of Chronic Pain
Division of Pain Medicine Assistant Attending, Department of Anesthesiology
Mayo Clinic and Critical Care
Rochester, Minnesota Memorial Sloan Kettering Cancer Center
New York, New York
Gregory Carpenter, MD, MBA
Department of Anesthesiology Robert J. Hackworth, MD
Vanderbilt University Division of Pain Medicine
Nashville, Tennessee Naval Medical Center
San Diego, California
John Corey, MD
Assistant Professor of Clinical Anesthesiology M. Gabriel Hillegass, III, MD
Division of Pain Medicine Assistant Professor of Anesthesiology
Vanderbilt University Division of Pain Medicine
Nashville, Tennessee Department of Anesthesia and Perioperative Medicine
Medical University of South Carolina
Bryan Covert, MD Charleston, South Carolina
Pain Medicine Fellow
Department of Anesthesiology Joshua Horowitz, DO
Vanderbilt University Pain Medicine Fellow
Nashville, Tennessee Department of Anesthesiology
Vanderbilt University
Kurt F. Dittrich, MD Nashville, Tennessee
Assistant Professor of Clinical Anesthesiology
Division of Pain Medicine Joseph C. Hung, M.D.
Vanderbilt University Interventional Pain Management Physician
Nashville, Tennessee Tripler Army Medical Center
United States Department of Defense
New York, New York

xi
Elizabeth Huntoon, MS, MD Ramana K. Naidu, MD
Assistant Professor Physical Medicine and Rehabilitation Assistant Professor of Anesthesia & Perioperative Care
Assistant Professor Orthopedic Surgery and Rehabilitation University of California, San Francisco
Director of Physical Medicine and Rehabilitation Medical San Francisco, California
Student Education
Vanderbilt Medical Group Andrea L. Nicol, MD, MS
Nashville, Tennessee Assistant Professor of Anesthesiology
University of Kansas School of Medicine
Adam K. Jacob, MD Kansas City, Kansas
Associate Professor of Anesthesiology
Mayo Clinic Ryan Nobles, MD
Rochester, Minnesota Assistant Professor of Anesthesiology
Division of Pain Medicine
Usman Latif, MD, MBA Department of Anesthesia and Perioperative Medicine
Assistant Professor of Anesthesiology Medical University of South Carolina
University of Kansas School of Medicine Charleston, South Carolina
Kansas City, Kansas
Daniel Pak, MD
Daniel F. Lonergan, MD Resident
Assistant Professor of Clinical Anesthesiology Department of Anesthesiology
Division of Pain Medicine Weill Cornell Medical College
Vanderbilt University New York-Presbyterian Hospital
Nashville, Tennessee New York, NY

Maureen F. McClenahan, MD, CDR, MC, USN Meenal Patil, MD

Assistant Professor of Anesthesiology Pain Management Center
Uniformed Services University of the Health Sciences Trenton, New Jersey
Director of Pain Medicine, Department of Anesthesiology and
Pain Medicine Julie R. Price, PsyD
Naval Medical Center Co-​Interim Director for Health Psychology Services
Portsmouth, Virginia Osher Center for Integrative Medicine at Vanderbilt
Assistant Professor of Clinical Psychiatry
Kelly McQueen, MD, MPH Assistant Professor of Physical Medicine & Rehabilitation
Professor of Anesthesiology Vanderbilt University School of Medicine
Division of Ambulatory Anesthesiology Nashville, Tennessee
Director, Vanderbilt Anesthesia Global Health &
Development Micah J. Price, PsyD
Vanderbilt University Director, Department of Psychology
Nashville, Tennessee Broward Health Medical Center
Assistant Clinical Professor of Psychology
Neel Mehta, MD Coordinator of Internship Training and
Medical Director, Pain Medicine Liaison Services
Department of Anesthesiology Nova Southeastern University
Weill Cornell Medical College Fort Lauderdale, Florida
New York-Presbyterian Hospital
New York, NY Antonio Quidgley-​Nevares, MD
Associate Professor and Chairman of Physical Medicine &
Susan M. Moeschler, MD Rehabilitation
Department of Anesthesiology Staff Physiatrist and Pain Medicine Physician
Division of Pain Medicine Eastern Virginia Medical School
Mayo Clinic Norfolk, Virginia
Rochester, Minnesota

x ii • Con t r i bu tor s
Ellen W. K. Rosenquist, MD Drew M. Trainor, DO
Assistant Professor of Anesthesiology Denver Back Pain Specialists
Cleveland Clinic Lerner College of Medicine Greenwood Village, CO
Case Western Reserve University
Cleveland, Ohio Anthony A. Tucker, MD
Assistant Professor of Anesthesiology
Martha J. Smith, MD Uniformed Services University of the Health Sciences
Assistant Professor of Anesthesiology Staff Anesthesiologist and Pain Medicine Physician
Division of Pain Medicine Naval Medical Center
Vanderbilt University Portsmouth, Virginia
Nashville, Tennessee
Erik P. Voogd, MD
Christopher Sobey, MD Division of Pain Medicine
Assistant Professor of Clinical Naval Medical Center
Anesthesiology San Diego, California
Division of Pain Medicine
Vanderbilt University Jenna L. Walters, MD
Nashville, Tennessee Assistant Professor of Clinical Anesthesiology
Division of Pain Medicine
Natalie Strickland, MD Vanderbilt University
Assistant Professor of Anesthesiology Nashville, Tennessee
Emory University School of Medicine
Egleston Children’s Hospital Aaron Jay Yang, MD
Atlanta, Georgia Assistant Professor of Physical Medicine and Rehabilitation
Vanderbilt University
Hans P. Sviggum, MD Nashville, Tennessee
Medical Director, Obstetric Anesthesiology
Mayo Clinic Robert Yang, MD
Rochester, Minnesota Washington, DC

  C o n t r i b u t o r s • x i i i
 1.
PAIN A NATOMY A ND PHYSIOLOGY

Daniel J. Pak and Neel Mehta

I N T RODUC T ION D. C, Aδ, Aβ

E. Aδ, Aβ, C
This chapter focuses on pain anatomy and physiology to pro-
vide a comprehensive review of the mechanisms of nocicep- 3. The following are true statements regarding abdomi-
tion for preparation of the American Board of Anesthesiology nal visceral pain except:
Pain Medicine (PM) Examination. It reviews the anatomy
of pain pathways (particularly the spinothalamic sensory A. Nociceptive transmission occurs via C fibers.
tract) and the process of pain conduction from peripheral B. Distention of a hollow viscus has decreased intensity
nociceptors to the cerebral cortex. It also reviews the differ- of pain compared to gut transection.
ent mechanisms of sensitization and inhibition at peripheral C. Nociceptive transmission occurs with efferent
nociceptors (manifested as primary and secondary hyperal- sympathetic nerve fibers.
gesia), the spinal cord (wind-​up and sensitization of second-​ D. Most gastrointestinal pain is characterized as a dull,
order neurons), and supraspinal structures, which all affect aching sensation at the midline.
the processing of nociceptive signals in the nervous system E. Visceral pain is often associated with abnormal
and, ultimately, the perception of pain. sympathetic activity.

QU E S T IONS 4. Afferent pain fibers from most abdominal viscera

nociceptors travel with sympathetic fibers in which of
1. Which of the following statements about nociceptors the following?
is false?
A. Celiac plexus
A. Silent nociceptors respond to inflammation. B. Superior hypogastric plexus
B. Polymodal nociceptors are the most prevalent type. C. Ganglion impar
C. Mechanonociceptors respond to pinch and pinprick D. Stellate ganglion
sensations. E. Hepatic plexus
D. After repeated stimulation, nociceptors may
demonstrate sensitization. 5. All of the following statements are true regarding
E. Nociceptors have low thresholds for activation. ascending pain pathways except:

2. Arrange Aβ, Aδ, and C nerve fibers from the fastest to A. All first-​order afferent nerve fibers enter the spinal
slowest conduction velocities. cord via the dorsal spinal root.
B. First-​order neurons may synapse with sympathetic
A. Aβ, C, Aδ neurons.
B. Aβ, Aδ, C C. Second-​order neurons in the dorsal horn mostly
C. C, Aβ, Aδ decussate to the contralateral side.

1
 D. Wide dynamic range (WDR) neurons are second-​ house fire. On physical exam, he has intense pain after
order neurons. application of mild heat at the site of injury. This is an
E. Third-​order neurons are located in the thalamus and ­example of:
send nerve fibers to the cortex.
A. Primary hyperalgesia
B. Secondary hyperalgesia
6. Which of the following are excitatory neurotransmit-
C. Allodynia
ters that modulate pain?
D. Paresthesia
A. Substance P, glutamate, aspartate, γ-​aminobutyric E. Disinhibited pain
acid (GABA)
B. Substance P, glutamate, enkephalins, serotonin 11. The same patient (from Question 10) returns 1 year
C. Glutamate, aspartate, substance P, adenosine later in clinic and says that after exposure to cold tem-
triphosphate (ATP) peratures, he now experiences burning sensations in
D. Glutamate, serotonin, GABA, enkephalins areas that were not injured during the house fire. Which
E. GABA, enkephalins, serotonin, ATP of the following statements is incorrect regarding the
patient’s symptoms and central hyper­sensitivity?
7. Release of substance P may cause all of the following
except: A. WDR neurons exhibit a reduction in neural
activation thresholds.
A. Sensitization of nociceptors B. WDR. neurons have increased frequency of
B. Vasoconstriction discharge with the same stimuli.
C. Enhanced chemotaxis C. Substance P increases sensitization.
D. Mast cell degranulation D. Nonsteroidal anti-​inflammatory drugs (NSAIDs) do
E. Increased neurokinin-​1 activity not have analgesic action on the spinal cord.
E. NMDA receptor activation increases sensitization.
8. All of the following statements correctly characterize
GABA activity except:
12. A 71-​year-​old man who underwent an exploratory
A. GABA is similar to glycine in that both are laparotomy for small bowel obstruction 2 weeks
inhibitory neurotransmitters that inhibit ascending ago now experiences increased pain and allodynia
pain pathways. at the surgical incision site. What direct role would
B. Inhibition of pain signal transmission is facilitated NSAIDs have on relieving this patient’s primary
by GABA B receptor activity. hyperalgesia?
C. GABA is an antagonist for N-​methyl-​d-​aspartate
(NMDA) receptors. A. Decrease prostacyclin and prostaglandin release
D. Benzodiazepines act on GABA A receptors. B. Decrease serotonin release
E. All of the above statements are true. C. Decrease substance P release
D. Decrease calcitonin gene-​related peptide (CGRP)
release
9. Which of the following statements is true regarding E. Decrease all of the above substances
WDR neurons compared to second-​order nociceptive-​
specific neurons?
13. Which of the following statements is true regarding
A. WDR neurons have smaller receptive fields the spinothalamic tract?
compared to nociceptive-​specific neurons.
B. WDR neurons are more prevalent than nociceptive-​ A. The lateral spinothalamic tract projects mainly to
specific neurons. the ventral posterolateral nucleus of the thalamus.
C. WDR neurons decrease their firing rate with B. The lateral spinothalamic tract mediates emotional
repeated stimulation. pain perception.
D. WDR neurons only respond to noxious input. C. The medial spinothalamic tract carries information
E. Nociceptive-​specific neurons can respond to non-​ regarding intensity and location of pain.
noxious stimuli. D. The medial spinothalamic tract mediates perceptions
of vibration and proprioception.
10. A 37-​year-​old otherwise healthy male experiences E. The spinothalamic tract mainly ascends in the gray
second-​degree burns on the right forearm following a matter of the spinal cord.

2 • Pa in M edicine B oa rd R e v ie w
14. The periaqueductal gray (PAG) produces analgesia D. Decreased release of GABA
by all of the following mechanisms except: E. Lack of endogenous opiate release

A. Activation of interneurons in lamina II 18. The patient from Question 17 asks you about trans-
B. Release of endogenous opioids cutaneous electrical nerve stimulation (TENS) as an
C. Inhibition of first-​order neurons additional modality of treatment for her pain. What
D. Decreased release of substance P is the most likely reason for the efficacy of TENS with
E. All of the above statements are true relation to the gate control theory?
15. A 65-​year-​old male with intra-​abdominal and retro- A. Inhibition of Aβ nerve fibers
peritoneal masses has persistent left-​sided abdominal B. Inhibition of the PAG
pain despite noninterventional treatments. His pain C. Inhibition of cutaneous nociceptors
was refractory to oral medical therapy as well as intra- D. Inhibition of C pain fibers
thecal opioids. He underwent a unilateral percutaneous E. Inhibition of B fibers
cervical cordotomy, an ablation procedure of the lateral
spinothalamic (neospinothalamic) tract. However, he
19. A 49-​year-​old male with a history of poorly control­
continues to complain of pain and severe distress fol-
led type 2 diabetes mellitus presents to the pain clinic
lowing treatment. Which of the following statements is
with bilateral lower extremity pain that is described as a
most likely true regarding this patient?
constant burning sensation. He would like to avoid any
opioid use due to gastrointestinal intolerances. He asks
A. Cordotomies are more effective for relieving central
about the possible use of amitriptyline, a tricyclic anti-
pain rather than peripheral pain syndromes.
depressant, as an analgesic option. Which of the follow-
B. Ablation procedures such as cordotomies should
ing would be an appropriate response?
not be recommended for patients with poor life
expectancies.
A. Tricyclic antidepressants would not be helpful
C. A bilateral cordotomy would have been more because they do not have analgesic properties.
effective for this patient. B. Tricyclic antidepressants would not be helpful
D. Ablation of the medial spinothalamic because they are only effective for nociceptive pain
(paleospinothalamic) tract could also be considered syndromes.
for this patient.
C. Tricyclic antidepressants are adequate alternative
E. Cordotomies are not effective for reducing analgesics because they predominantly act on μ-​
nociceptive pain. opiate receptors.
16. Which of the following statements is most likely D. Tricyclic antidepressants are adequate alternative
correct regarding the previous patient’s pain (from analgesics because they predominantly increase
Question 15)? supraspinal inhibition.
E. Tricyclic antidepressants are adequate alternative
A. Inhibition of descending pain modulating pathways analgesics because they predominantly act as glycine
should decrease this patient’s pain. agonists.
B. Spinothalamic fibers do not project to the PAG in
the midbrain. 20. Spinal cord gray matter is composed of 10 layers called
C. The PAG stimulates ascending pain fibers. Rexed laminae. Aδ fibers synapse predominantly on:
D. All pain fibers decussate to the contralateral
spinal cord. A. Lamina I and II
E. Other ascending pain pathways may contribute to B. Lamina I and V
this patient’s pain. C. Lamina III and IV
D. Lamina VII
17. A 58-​ year-​old woman with recently diagnosed E. Lamina II and V
postherpetic neuralgia presents to the pain clinic after
complaining of persistent hyperesthesia and allodynia. 21. Neurons in Rexed lamina II (substantia gelatinosa)
Which of the following best describes the most likely play an important role in modulating pain perception
reason for her increased pain? due to:

A. Decreased glutamate release A. Inhibitory projections to ascending

B. Activation of NMDA receptors spinothalamic fibers
C. Lack of oral narcotic use B. Increased release of substance P

1. Pa in An atom y a nd Ph y sio l o g y • 3
 C. Decreased release of GABA D. Postsynaptic opiate receptor activation causes
D. High levels of WDR neurons that increase neuronal hyperpolarization.
responsiveness to only noxious stimuli E. Endogenous opiate peptides are antagonized by
E. Increased sensitization of cutaneous nociceptors naloxone.
23. A 68-​year-​old male patient with a history of stage
22. All of the following are true regarding the actions IVB hepatocellular carcinoma and diffuse bone metas-
of endogenous opiates except: tases whom you have been treating for intractable pain
has a pathologic femoral bone fracture. Which of the
A. Opioid receptors are distributed widely throughout following systemic responses would you expect to see in
the central nervous system, including the cerebral this patient?
cortex, brainstem, dorsal horn, and dorsal root
ganglion. A. Tachycardia
B. PAG stimulation decreases endogenous opiate B. Hyperglycemia
release. C. Increased oxygen consumption
C. Presynaptic opiate receptor activation inhibits the D. Hypercoagulability
release of glutamate and substance P. E. All of the above

4 • Pa in M edicine B oa rd R e v ie w
 A NSW ER S F U RT H E R R E A DI NG

1. ANSWER: E Barash PG, Cullen BF, Stoelting RK, et al. Clinical Anesthesia. 7th ed.
Philadelphia, PA: Lippincott Williams & Wilkins; 2013.
The vast majority of nociceptors are free nerve endings that
respond to noxious stimuli, such as heat, mechanical, and
chemical tissue injury. They have high activation thresh-
olds and can increase neural firing rates depending on the 3. ANSWER: B
intensity of the stimuli. Nociceptors also demonstrate func-
tional plasticity, and with repeated stimulation they have Visceral nociceptors respond to chemical or mechanical
sensitization. This can cause nerve transmission following stimuli, such as distention, ischemia, and inflammation.
a low-​intensity noxious stimulus (hyperalgesia) or even after They do not respond as intensely to the localized transec-
a non-​noxious stimulus (allodynia). The three major noci- tion or burning associated with surgery. These nerve fibers
ceptor types are mechanonociceptors (responsive to pinch also synapse at several dermatomal levels and can cross the
and pinprick sensations), silent nociceptors (responsive to contralateral dorsal horn. As a result, pain is usually per-
inflammation), and polymodal mechano-​heat nociceptors ceived at the midline and is characterized as a nonspecific
(the most common type; responsive to pressure, tempera- dull and aching sensation. Visceral afferents travel via unmy-
ture, and neurochemical mediators such as histamine, cap- elinated C fibers alongside efferent sympathetic nerve fibers.
saicin, and bradykinin). Consequently, afferent activity from these nociceptors is
transmitted to the spinal cord between the levels of T1 and
L2, and pain can be associated with abnormal sympathetic
F U RT H E R R E A DI NG activity such as nausea, vomiting, and hemodynamic shifts.

Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s

Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013. F U RT H E R R E A DI NG

Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s

Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.

2. ANSWER: B

Aβ afferent nerve fibers transmit non-​ noxious stim- 4. ANSWER: A

uli, whereas Aδ and C fibers transmit noxious stimuli
(Table 1.1). In general, nerve conduction velocity is The celiac plexus, which is anterior to the vertebral body of
dependent on the diameter and degree of myelination of L1, carries afferent nociceptive and sympathetic innervation
the nerve axon. Myelination increases the electrical insu- for most of the abdominal viscera. This includes innerva-
lation of the nerve and conducts the impulse at a higher tion for the liver, pancreas, biliary tract, gallbladder, adrenal
velocity due to salutatory conduction along the axon. glands, spleen, kidneys, stomach, and small and large bow-
Larger nerves also have improved electrical conduction. els. Therefore, celiac plexus blocks are effective for reduc-
As a result, conduction velocities are fastest for large, ing chronic pain from upper abdominal visceral pathology
myelinated A fibers compared to smaller, nonmyelinated (i.e., intractable pain from pancreatic cancer).
C fibers. The superior hypogastric plexus, which is located ante-
riorly to the L5–​S2 vertebral bodies, carries afferent and
sympathetic innervation for the large bowel distal to the left
colonic flexure and the pelvic viscera. The ganglion impar,
Table 1.1 CHAR ACTERISTICS OF MAJOR AFFER ENT
which is located on the anterior surface of the coccyx, also
NERVE FIBERS
provides afferent and sympathetic innervation for pelvic
C FIBERS AΔ FIBERS AΒ FIBERS viscera. The stellate ganglion, located anteriorly to the C7
vertebral body, carries afferent and sympathetic innervation
Function Diffuse, Sharp, localized Light touch
dull pain pain to portions of ipsilateral head, neck, and arm.
Size (diameter; (0.3–​1.6) (1–​4) (6–​22)
μm) F U RT H E R R E A DI NG
Myelination –​ + ++
Conduction <2 5–​25 >30 Kambadakone A, Thabet A, Gervais DA, et al. CT-​g uided celiac plexus
velocity (m/​s) neurolysis: A review of anatomy, indications, technique, and tips for
successful treatment. RadioGraphics. 2011;31(6):1599–​1621.

1. Pa in An atom y a nd Ph y sio l o g y • 5
5. ANSWER: A (NK-​1) receptors to facilitate pain transmission. Other
important excitatory neurotransmitters are glutamate and
The transmission of pain signals starts with activation of aspartate (both act on NMDA receptors), CGRP, and ATP.
first-​order neurons, which mostly enter the dorsal horn of
the spinal cord from the periphery via the dorsal spinal root.
A minority of first-​order neurons, however, may enter the spi- F U RT H E R R E A DI NG
nal cord via the ventral nerve root, which is why some patients
who undergo rhizotomies (transection of dorsal nerve roots Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
in chronic pain patients for analgesia) can continue to feel
pain following the procedure. First-​order neurons may travel
up or down several spinal segments in Lissauer’s tract prior
to synapsing with second-​order neurons in the dorsal horn,
which mostly decussate and cross the midline to the contra- 7. ANSWER: B
lateral side of the spinal cord to ascend in the spinothalamic
tract. Second-​order neurons are either nociceptive-​specific or Substance P is a neuropeptide that plays a major role as an
WDR neurons. Both types receive noxious input from Aδ excitatory neurotransmitter in nociception through its acti-
and C fibers, but WDR neurons receive non-​noxious input vation of NK-​1 receptors. It is synthesized and released in
as well. Second-​order neurons then synapse with third-​order response to painful stimuli from peripheral terminals of
neurons in the thalamus, which send fibers through the sensory nerve fibers and first-​order neurons in the dorsal
internal capsule to the cerebral cortex. horn. In addition to its role as a facilitator of pain path-
ways, substance P sensitizes nociceptors, causes histamine
degranulation from mast cells, and causes 5-​HT release
F U RT H E R R E A DI NG from platelets. Substance P is also a potent vasodilator via
its activity on NK-​1 receptors on the endothelium of blood
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s vessels.
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
Hemmings HC, Egan TD. Pharmacology and Physiology for
Anesthesia: Foundations and Clinical Application. Philadelphia,
PA: Elsevier; 2013.
F U RT H E R R E A DI NG

Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s

Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
6. ANSWER: C

There are both excitatory and inhibitory neurotransmitters

that act on afferent neurons transmitting pain informa- 8. ANSWER: C
tion (Table 1.2). Substance P is an excitatory neuropeptide
released by first-​order neurons that acts on neurokinin-​1 Both GABA and glycine are important neurotransmitters
that are released by inhibitory interneurons in the dorsal
Table 1.2 NEUROTR ANSMITTERS IN PAIN horn and have an essential role for inhibiting other excit-
MODULATION atory neural pathways. The major neurotransmitter for
excitatory interneurons is glutamate. Through inhibition of
NEUROTR ANSMITTER MODULATING EFFECT WDR neurons and ascending pain fibers of the spinotha-
Glutamate Excitatory lamic tract in the dorsal horn, G protein-​coupled GABA B
Aspartate Excitatory receptor activity plays an important role for analgesia. In
fact, inhibition of GABA B receptors can lead to hyperes-
Substance P Excitatory
thesia and allodynia. GABA A receptors, on the other hand,
Calcitonin gene-​related Excitatory
peptide (CGRP)
are ligand-​gated ion channels that increase Cl–​ influx.
Drugs such as benzodiazepines and barbiturates act on
Adenosine triphosphate (ATP) Excitatory
these receptors. GABA does not have any inherent NMDA
γ-​Aminobutyric acid (GABA) Inhibitory
receptor activity.
Acetylcholine Inhibitory
Enkephalins Inhibitory
β-​Endorphins Inhibitory F U RT H E R R E A DI NG
Serotonin Inhibitory
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Norepinephrine Inhibitory
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.

6 • Pa in M edicine B oa rd R e v ie w
Hemmings HC, Egan TD. Pharmacology and Physiology for and decreases the threshold for neural transmission. An
Anesthesia: Foundations and Clinical Application. Philadelphia, enhanced response to the same stimulus intensity can be
PA: Elsevier; 2013.
demonstrated, and continued transmission of pain signals
following resolution of the stimulus is also common.
Secondary hyperalgesia develops in the region immedi-
ately surrounding the injured tissue and is a result of sub-
9. ANSWER: B
stance P release, which causes tissue edema, reddening or
flaring of the skin around the site of injury, and sensitiza-
First-​order neurons in the ascending pain pathway synapse
tion to noxious stimuli. Unlike primary hyperalgesia, which
with either second-​order nociceptive-​specific neurons or
occurs in response to both mechanical and heat stimuli, sec-
WDR neurons in the dorsal horn. WDR neurons are the
ondary hyperalgesia is triggered only by mechanical stimuli.
most abundant in the dorsal horn. Both nociceptive-​spe-
Allodynia refers to perception of pain following a non-​
cific and WDR neurons receive noxious input from Aδ and
noxious stimulant, whereas paresthesia is an abnormal sen-
C fibers, but WDR neurons receive non-​noxious input as
sation (usually tingling or pricking) without an apparent
well. Therefore, WDR neurons are characterized by large
stimulant.
receptive fields, whereas nociceptive-​specific neurons have
smaller, discrete receptive fields that are normally silent and
responsive only to high-​threshold noxious input. WDR
F U RT H E R R E A DI NG
neurons play a key role in central sensitization of pain.
Repeated stimulation can exponentially increase the rate of Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
firing of WDR neurons, causing a “wind-​up” phenomenon Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
that leads to increased second-​order pain transduction for
the same stimulus intensity.

11. ANSWER: D
F U RT H E R R E A DI NG
Central sensitization refers to the enhancement and
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s decreased inhibition of nociceptive pain pathways in the
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013. spinal cord to produce pain hypersensitivity. This can occur
Hemmings HC, Egan TD. Pharmacology and Physiology for
Anesthesia: Foundations and Clinical Application. Philadelphia, following an intense noxious stimuli or repeated exposure
PA: Elsevier; 2013. to stimuli.
Facilitation of central sensitization occurs primarily
through one or more of the following mechanisms: (1) sen-
sitization of second-​order neurons where WDR neurons
10. ANSWER: A have increased response and frequency of discharge fol-
lowing stimulation, (2) a reduction in the neural activation
This patient is experiencing hyperalgesia, which is an threshold, and (3) an enlargement of the receptor fields such
enhanced response to a noxious stimulant (Figure 1.1). that adjacent neurons in the dorsal horn become responsive
Injury to the skin can cause two types: primary and sec- to both noxious (hyperalgesia) and innocuous (allodynia)
ondary hyperalgesia. Primary hyperalgesia occurs at the site stimuli. This patient is experiencing allodynia in areas
of injury and is due to release of various chemical modu- beyond the original site of injury.
lators by the injured tissue (histamine, serotonin, brady- Substance P is one of the main mediators of central
kinin, and prostaglandins). This sensitizes nociceptors sensitization by facilitating increased neural membrane
excitability through its interaction with G protein-​
coupled membrane receptors. Excitatory amino acids
Hyperalgesia
such as glutamate and aspartate also facilitate central
sensitization through its activity on NMDA receptors.
Prostaglandins help activate the release of these amino
Perception acids in the spinal cord; therefore, NSAIDs help reduce
of pain central sensitization.

F U RT H E R R E A DI NG
Intensity of stimulant
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Figure 1.1 Hyperalgesia effect of pain perception. Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.

1. Pa in An atom y a nd Ph y sio l o g y • 7
Latremoliere A, Woolf CJ. Central sensitization: A generator Hemmings HC, Egan TD. Pharmacology and Physiology for
of pain hypersensitivity by central neural plasticity. J Pain. Anesthesia: Foundations and Clinical Application. Philadelphia,
2009;10(9):895–​926. PA: Elsevier; 2013.

12. ANSWER: A 14. ANSWER: E

Tissue injury leads to the release of inflammatory media- The PAG of the midbrain plays an important role in supra-
tors. This includes the production and release of prostaglan- spinal inhibition of ascending pain afferents. Stimulation
dins, including prostaglandin E2 (PGE2), which activate of the PAG promotes excitatory connections with inhibi-
and sensitize nociceptors. This leads to decreased nocicep- tory interneurons in Rexed lamina II of the dorsal horn,
tor threshold for firing and an increased response to nox- which release endogenous opioids, such as enkephalin,
ious stimuli as seen with primary hyperalgesia. NSAIDs that bind to μ-​opiate receptors on axons of Aδ and C nerve
counteract primary hyperalgesia through the decreased fibers. Opiate receptor activation subsequently decreases
production of prostacyclin and prostaglandins via inhibi- substance P release from these primary afferent neurons,
tion of the cyclooxygenase (COX) pathway. Recent studies thereby inhibiting further activation of ascending second-​
have also indicated that COX inhibitors may decrease cen- order neurons and transmission of ascending pain signals.
tral sensitization of nociception in the spinal cord as well. The PAG can also evoke antinociceptive action through
Neurochemical mediators such as substance P, CGRP, and adrenergic α2 receptor activation in the dorsal horn. Deep
serotonin all have important excitatory effects on nocicep- brain stimulation of the PAG has been demonstrated to
tion but are not directly affected by NSAIDs. provide pain relief for some intractable pain syndromes,
but it is not widely employed and remains “off-​label” in the
United States.
F U RT H E R R E A DI NG

Sinatra RS, Leon-​Cassasola OA de, Viscusi ER. Acute Pain Management.

New York, NY: Cambridge University Press; 2009. F U RT H E R R E A DI NG

Boccard SGJ, Pereira EAC, Aziz TZ. Deep brain stimulation for
chronic pain. J Clin Neurosci. 2015;22(10):1537–​1543.
Budai D, Harasawa I, Fields HL. Midbrain periaqueductal gray
13. ANSWER: A (PAG) inhibits nociceptive inputs to sacral dorsal horn nocicep-
tive neurons through α2-​adrenergic receptors. J Neurophysiol.
1998;80(5):2244–​2254.
The spinothalamic tract is the major ascending pain path-
way that travels in the anterolateral portion of the spinal
cord white matter. Axons of second-​order neurons decus-
sate via the anterior white commissure and ascend on the
contralateral side to eventually project to supraspinal struc- 15. ANSWER: D
tures such as the thalamus, nucleus raphe magnus, and
periaqueductal gray. Second-​order neurons then synapse Patients with malignancies can experience severe pain that
with third-​order neurons in the thalamus to eventually is refractory to opiates and intrathecal infusion therapies.
project to the primary somatosensory cortex and the cin- Therefore, ablation procedures of the spinothalamic tract
gulate gyrus. It consists of two main pathways: the medial can be performed to relieve persistent nociceptive pain, with
(paleospinothalamic) and lateral (neospinothalamic) tracts. percutaneous cordotomy being an effective option for those
The neospinothalamic tract transmits information regard- with refractory unilateral nociceptive pain. Although effec-
ing the location, duration, and intensity of pain and com- tive at reducing pain, ablation procedures are usually rec-
municates to the ventral posterolateral nucleus (VPN) of ommended only in patients with limited life expectancies
the thalamus. The paleospinothalamic tract transmits emo- because the analgesic effects diminish over time. It is also
tional perceptions of pain and communicates to the medial more effective at relieving pain from peripheral nociceptor
thalamus. activation as seen with malignant pain syndromes in which
tumors infiltrate bones or nerves. Neuropathic and central
pain syndromes have less predictable results. Bilateral abla-
F U RT H E R R E A DI NG tions are not commonly performed and are indicated only
for bilateral or midline pain.
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s Cordotomies typically target ascending pathways in
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013. the lateral spinothalamic (neospinothalamic) tract, which

8 • Pa in M edicine B oa rd R e v ie w
transmits information regarding location, duration, and Palecek J, Paleckova V, Willis WD. The roles of pathways in the spinal
intensity of pain. Ablations of the medial spinothalamic cord lateral and dorsal funiculi in signaling nociceptive somatic and
visceral stimuli in rats. Pain. 2002;96(3):297–​307.
(paleospinothalamic) tract do not provide analgesia by
usual pain metrics, but they can provide relief of distress in
malignancy pain syndromes and be helpful for managing
the emotional perceptions of pain. Therefore, following an
17. ANSWER: B
ablation of the paleospinothalamic tract, this patient may
continue to perceive pain but not be distressed by it. Note
This patient is experiencing wind-​up, a mechanism of cen-
that for this particular question, it is not necessary to know
tral pain sensitization. With repeated activation of C fibers,
the specifics of what a cordotomy entails. However, one
a progressive increase in the evoked response is seen such
should know the differences between the two major spino-
that WDR neurons increase their frequency of firing and
thalamic pathways.
also have prolonged firing after resolution of the stimulus.
Glutamate and aspartate, both excitatory neurotransmitters,
are important facilitators for wind-​up through their activa-
F U RT H E R R E A DI NG tion of NMDA receptors, which can be found on WDR
Frost EAM. Clinical Anesthesia in Neurosurgery. 2nd ed. Boston, MA:
neurons. NMDA antagonists, such as ketamine, have been
Butterworth-​Heinemann; 1991. found to reduce wind-​up pain in patients with postherpetic
Tollison CD, Satterthwaite JR, Tollison JW. Practical Pain neuralgia and central neuropathic pain syndromes as well.
Management. 3rd ed. Philadelphia, PA: Lippincott Williams &
Wilkins; 2002.
F U RT H E R R E A DI NG

Eide PK. Wind-​up and the NMDA receptor complex from a clinical
16. ANSWER: E perspective. Eur J Pain. 2000;4(1):5–​15.

Because there are multiple ascending pain pathways, some

patients may continue to perceive pain following ablation
procedures of the spinothalamic tract. Like the spinotha- 18. ANSWER: D
lamic tract, the spinoreticular tract fibers decussate and
ascend in the contralateral spinal cord to transmit signals TENS is a noninvasive treatment that applies electrical cur-
to the thalamus and hypothalamus. These fibers mediate rent to the skin. This tactile, non-​noxious stimulation acti-
the emotional and autonomic aspects of pain. The spino- vates Aβ fibers and modulates afferent information carried
cervical tract, another pathway for transmission of noci- by ascending C pain fibers by inhibition within the sub-
ceptive pain, ascends ipsilaterally to the lateral cervical stantia gelatinosa (or Rexed lamina II of the dorsal horn).
nucleus and projects to the contralateral thalamus. Dorsal This descending inhibition initiated by Aβ fibers is achieved
column fibers that travel ipsilaterally and that were tra- via stimulation of the PAG. This has often been referred to
ditionally thought to transmit signals for proprioception as the gate control theory of pain. TENS has been associ-
have also been implicated to transmit visceral nociceptive ated with improving pain symptoms and decreasing opiate
information. requirements in neuropathic pain syndromes, including
The PAG in the midbrain and the rostral ventromedial postherpetic neuralgia, as well as in other acute pain
medulla are involved in regulating descending pathways syndromes.
that project to the dorsal horn to inhibit the ascending pain
tracts. Therefore, stimulation of these areas with implants
has been demonstrated to be potentially helpful in patients F U RT H E R R E A DI NG
with chronic pain. It is important to know that fibers of the
spinothalamic tract also project to the PAG and can modu- Breivik H, Campbell WI, Nicholas MK. Clinical Pain
late descending inhibitory pathways. Management: Practice and Procedures. 2nd ed. Boca Raton,
FL: CRC Press; 2008.

F U RT H E R R E A DI NG
19. ANSWER: D
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
Frost EAM. Clinical Anesthesia in Neurosurgery. 2nd ed. Boston, Unlike nociceptive pain syndromes, chronic neuropathic
MA: Butterworth-​Heinemann; 1991. pain syndromes are due to primary damage to nerve fibers

1. Pa in An atom y a nd Ph y sio l o g y • 9
of the peripheral or central nervous system. These condi- relay information regarding touch and proprioception.
tions can be difficult to treat with conventional analgesics Lamina V processes visceral and somatic pain as well as
such as opiates and NSAIDs alone. Antidepressants such as non-​noxious afferent information. Lamina VI relays pro-
tricyclic antidepressants (TCAs) and anticonvulsants such prioception information. Lamina VII contains pregangli-
as gabapentin or pregabalin are typically considered first-​ onic sympathetic neurons. Lamina VIII and IX comprise
line agents. TCAs inhibit serotonin and catecholamine the anterior motor horn. Nociceptive C fibers synapse pre-
reuptake in the synaptic nerve cleft, which increases mono- dominantly with second-​order neurons in laminae I and II,
amine-​mediated inhibition of ascending pain pathways, whereas Aδ fibers terminate mainly on laminae I and V.
thereby producing analgesia. This patient is experiencing
diabetic neuropathic pain from his poorly controlled dis-
ease and would be considered an excellent candidate for F U RT H E R R E A DI NG
antidepressant therapy.
Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
McMahon SB, Koltzenburg M, Tracey I, et al. Wall and Melzack’s
F U RT H E R R E A DI NG Textbook of Pain. 6th ed. Philadelphia, PA: Elsevier, 2013.

Moore RA, Derry S, Aldington D, et al. Amitriptyline for neuropathic

pain in adults. Cochrane Database Syst Rev. 2015;7:CD008242.

21. ANSWER: A

20. ANSWER: B Rexed lamina II (substantia gelatinosa) neurons receive

input from both Aδ and C fibers and play an essential role
The spinal cord gray matter is divided into 10 Rexed laminae in modulating spinothalamic nerve fibers through their
(Table 1.3), which follow a topographic organization and inhibitory interneurons. They do not release substance
are organized based on different functions. Laminae I–​V I P. Lamina II interneurons also play an important role in
comprise the dorsal horn and serve as the major area where analgesia through their expression of μ-​opioid receptors.
both ascending and descending spinal pathways modulate Although WDR neurons are common in the dorsal horn,
pain. Some of the important layers are discussed here. they are most prevalent in lamina V. These second-​order
neurons have excitatory projections and respond to both
Table 1.3 R EXED LAMINAE FUNCTIONS non-​noxious and noxious stimuli.

R EXED LAMINAE FUNCTIONS

I Somatic nociception, thermoreception F U RT H E R R E A DI NG

II Somatic nociception, thermoreception,
opiate responsive Trafton JA, Abbadie C, Marek K, et al. Postsynaptic signaling
via the [mu] opioid receptor: Responses of dorsal horn neu-
III Mechanoreception, proprioception rons to exogenous opioids and noxious stimulation. J Neurosci.
2000;20(23):8578–​8584.
IV Mechanoreception, proprioception
V Visceral, somatic nociception,
mechanoreception
VI Mechanoreception, proprioception 22. ANSWER: B
VII Preganglionic sympathies
The endogenous opiate system consists of three main pep-
VIII Anterior motor horn tides (β-​endorphin, enkephalins, and dynorphins) and
IX Anterior motor horn three main G protein-​coupled receptors (μ, δ, and κ), which
are widely expressed in the central nervous system, includ-
X Central gray commissures
ing the cerebral cortex, brainstem, limbic system, dorsal
horn, and dorsal root ganglion. The release of endogenous
Lamina I contains second-​order neurons that receive opiates is triggered by activation of the PAG. Activation
cutaneous and deep somatic nociceptive pain and tempera- of opioid receptors leads to presynaptic inhibition of the
ture afferents. Lamina II (or substantia gelatinosa) contains release of excitatory chemical neurotransmitters, such as
second-​order neurons that modulate cutaneous nociceptive glutamate, substance P, and CGRP. Concurrently, opioid
information and are opiate responsive. Lamina III and IV receptor activation also causes postsynaptic hyperpolariza-
second-​order neurons receive non-​nociceptive input and tion for decreased neuronal excitability. Exogenous opioids

10 • Pa in M edicine B oa rd R e v ie w
have a predilection to act on second-​order neurons in the patients with preserved cardiac function. This may also lead
substantia gelatinosa of the spinal cord. Both endogenous to increased oxygen consumption with increased work of
and exogenous opiates are antagonized by naloxone. breathing. An increase in the release of catecholamines and
cortisol additionally leads to hyperglycemia as well as stress-​
related immunosuppression. Increased sympathetic activity
F U RT H E R R E A DI NG can cause urinary retention or ileus. Hypercoagulability
can also be seen with decreased fibrinolytic states. The
Benarroch EE. Endogenous opioid systems: Current concepts and clini- patient described in this question is experiencing severe
cal correlations. Neurology. 2012;79(8):807–​814.
acute, superimposed on chronic pain, which places him at a
high likelihood of developing the described systemic stress
response.
23. ANSWER: E
F U RT H E R R E A DI NG
A neuroendocrine stress response is seen with acute pain
syndromes that can be attributed to increased sympathetic Barash PG, Cullen BF, Stoelting RK, et al. Clinical Anesthesia. 7th ed.
activation and release of stress hormones. It can also be Philadelphia, PA: Lippincott Williams & Wilkins; 2013.
witnessed in chronic pain patients who experience promi- Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail’s
Clinical Anesthesiology. 5th ed. New York, NY: McGraw-​Hill; 2013.
nent recurring nociceptive and central pain syndromes. Hemmings HC, Egan TD. Pharmacology and Physiology for
Common cardiovascular effects include hypertension and Anesthesia: Foundations and Clinical Application. Philadelphia,
tachycardia, which lead to increased cardiac output for PA: Elsevier; 2013.

1. Pa in An atom y a nd Ph y sio l o g y • 11
 2.
LITER ATUR E R EV IEW A ND EV IDENCE

Andrea L. Nicol and Usman Latif

I N T RODUC T ION D. A common standard for levels of evidence allows

for a uniform approach to comparison of different
This chapter reviews concepts underlying critical analysis of sources of data.
literature and evidence-​based medicine because a thorough E. Ratings are based on the design and quality of the
understanding of these topics is of utmost importance in study or paper.
the interpretation of medical literature and applicability of
the results therein. Basic principles of valid clinical research 3. Which of the following components of study design
and components of clinical trials are reviewed. The chapter eliminates confounding by baseline variables, removes
explores specific topics pertaining to the designing, report- investigator bias, and guarantees that statistical tests
ing, and interpreting of clinical studies about the treatment will have valid false-​positive error rates?
of pain. The effects of the analysis on the clinical applicabil-
ity of study results are also discussed. Finally, the chapter A. Blinding
identifies special features specific to the study of pain. B. Sample size calculation
C. Informed consent
D. Randomization
QU E S T IONS E. Effect size

1. An investigator is performing a study in which there 4. A researcher is interested in performing a random-

will be 100 separate independent comparisons in the ized placebo-​ controlled trial for patients undergo-
analysis. At a significance level of 0.05, how many false-​ ing lumbar transforaminal steroid injections. Which
positive findings are possible on the average based on of the following study parameters poses the greatest
chance alone? ethical challenge in the development and design of the
proposed study?
A. 1
B. 5 A. Placebo control
C. 10 B. Blinding
D. 25 C. Sample size
E. 100 D. Inclusion criteria
E. Interim analysis
2. Which of the following is false regarding levels of
evidence? 5. Which of the following is not an element of an
informed consent?
A. Evidence meeting the highest standard is rated 1a.
B. An individual randomized control trial (RCT) with A. A statement that participation is voluntary
a narrow confidence interval (CI) is rated as the B. A description of the benefits to the subject or others
highest level of evidence possible. expected from the research
C. An expert opinion is rated as a lower level of C. Contact information for the US Department of
evidence than a case series. Health and Human Services

12
 D. A description of the foreseeable risks and 10. Which of the following parameters provides clini-
discomforts to the subject cians with information on both the clinical signifi-
E. The expected duration of the subject’s participation cance and the statistical significance of an inferential
statistical test?
6. Which of the following is true regarding grades of
recommendation? A. p value
B. Effect size
A. Grades of recommendation are used to describe the C. Standard deviation
quality of the collection of evidence supporting an D. Confidence interval
assertion using a range of 1 to 5. E. Sample size
B. Grades range from A to D.
C. 1a is the highest grade of recommendation.
D. Whether a recommendation is based on 11. Which of the following is not a core outcome
extrapolation is irrelevant to grading. domain for chronic pain clinical trials as recommended
E. Grades of recommendation are ordered from lowest by the Initiative on Methods, Measurement, and Pain
to highest levels of evidence strength. Assessment in Clinical Trials (IMMPACT)?

A. Symptoms and adverse events

7. Which of the following scales or questionnaires would
B. Physical functioning
be most beneficial to a pain researcher who is inter-
C. Emotional functioning
ested in measuring physical functioning as a marker of
D. Ratings of global improvement
health-​related quality of life in a study of patients with
E. Health care utilization
chronic low back pain?

A. Visual numerical pain score 12. All the following are factors that should be con-
B. Short Form (SF)-​36 sidered when deciding whether you should use a para-
C. Visual analogue pain score metric or nonparametric statistical analysis approach
D. McGill Pain Questionnaire except:
E. Brief Pain Inventory
A. The shape of the data distribution
B. The type of the data being analyzed
8. A study was performed to evaluate the efficacy of a new
C. The assumption that samples are independent
neuropathic drug in the treatment of postherpetic neu-
D. The type of study design used
ralgia. In the initial phase, all patients received the study
E. The assumption that variances are homogeneous
drug, and outcome measures for changes in visual numer-
ical pain scores were assessed before and after the study
period. After the study period was complete, the research- 13. Match the following terms to the statements:
ers analyzed the data and selected only those patients who • Case–​control study

responded with a 30% or greater reduction in pain to con- • Cross-​sectional survey

tinue the study in a randomized placebo-​controlled trial. • Crossover design

What specific type of study design is described here? • Randomized controlled trial
• Cohort study
A. Crossover study design
B. Retrospective study design A. What is a controlled trial in which each subject has
C. Enrichment design both therapies at various points in time?
D. N-​of-​1 study design B. What design is best for studying the effect of an
E. Adaptive study design intervention?
C. What is a study design in which data are obtained
9. An RCT study design is more likely to result in all of from groups who have been exposed or not exposed
the following except: to a variable of interest?
D. Which study design is best for the study of the effect
A. Unbiased distribution of confounders of predictive risk factors on an outcome?
B. Facilitation of statistical analysis E. The prevalence of a disease or risk factor can be
C. Increased expense quantified best with which study design?
D. Decreased volunteer bias F. What is the only feasible study design for the study
E. Attrition bias based on group allocation of very rare disorders?

2 .  Liter ature R e v ie w a nd E v idence  • 13

14. A clinical trial is performed to evaluate the effective- C. Satisfaction rating (very satisfied, moderately
ness of a new drug for fibromyalgia in a population sam- satisfied, etc.)
ple of 120 patients. The mean decrease in pain scores D. Height (cm)
after treatment in the active treatment group is 4.5. The E. Gender (male or female)
mean decrease in average pain scores after treatment
in the placebo group is 2.8. Assuming a normal distri- 19. All of the following statements about research char-
bution of the data, which statistical test is the best to acteristics are true except:
utilize in comparing the mean change in average pain
scores between the two groups? A. In a double-​blinded study, both patients and
providers are unaware of the patients’ group
A. Analysis of variance (ANOVA) assignment.
B. Student’s t-​test B. Randomization is a process of selecting from a
C. Wilcoxon rank-​sum test group in a manner that makes equal distribution of
D. Pearson coefficient of correlation confounders likely.
E. Chi-​square test C. If the patients who are likely to volunteer for a study
are different than the general population, that is an
15. Which of the following was developed as a tool to example of confounding.
facilitate the complete and transparent reporting of D. Stratification is a strategy in which patients are
trials and aid in their critical appraisal and interpreta- intentionally divided by an important characteristic
tion in response to suboptimal and inadequate report- prior to randomization.
ing of results from randomized controlled trials? E. In a triple-​blinded study, patients, providers,
and another group (e.g., data analyzers or
A. Cochrane database support staff) are unaware of the patients’ group
B. PubMed assignment.
C. Consolidated Standards of Reporting Trials
(CONSORT) Statement 20. Which of the following is the probability of failing
D. Meta-​analysis to reject the null hypothesis when there is an association
E. EMBASE between predictor and outcome?

16. What magnitude of change in visual analogue scale A. Power

pain score is reported to be consistent with at least a B. p value
moderately clinically meaningful reduction to chronic C. Effect size
pain patients? D. α
E. β
A. ≥20%
B. ≥30% 21. An investigator is researching the efficacy of two
C. ≥40% different drugs for painful diabetic polyneuropathy.
D. ≥50% In the study, there are two separate treatment periods
E. ≥60% separated by a period of time in which they are admin-
istered no medication. In the first treatment period,
17. Which statistical principle is defined by the inclu- half of the group will be administered drug A and
sion of all patients for analysis in the groups to which half will be administered drug B. In the second treat-
they were assigned, regardless of protocol adherence? ment period, the groups will receive the drug they did
not receive in the first treatment period. What type
A. Intention-​to-​treat of study design has this investigator employed for his
B. Crossover research?
C. Per protocol
D. Bootstrapping A. Crossover
E. Subgroup analysis B. Enrichment
C. Dose-​finding
18. Which of the following could be classified as an D. Randomized controlled trial
ordinal variable? E. Adaptive

A. Color of eyes 22. Which of the following groups of people are not
B. Temperature (Celsius) considered to be vulnerable populations in research as

14 • Pa in M edicine B oa rd R e v ie w
defined by the US Department of Health and Human A. Sample size
Services? B. Inclusion criteria
C. Recruitment
A. Children and minors D. Sampling error
B. Cognitively impaired persons E. Internal validity
C. Cancer patients
D. Pregnant women 25. Match the following terms to the statements:
E. Prisoners • Positive predictive value
• Number needed to treat (NNT)
23. All of the following parameters are required to cal- • Power
culate sample size except: • Specificity
• Sensitivity
A. Effect size
B. Variability (standard deviation) A. The inverse of the absolute risk reduction (ARR) is
C. α equal to which measure?
D. β B. Which measure correlates with the proportion of
E. p value negatives that are correctly identified as such?
C. The number of true positives divided by the sum of
24. Which of the following components of clinical trial true positives and false positives is equal to?
design specifies and defines the main characteristics of D. Which measure quantifies the likelihood of
the sample population relevant to the research question? identifying a significant effect when it exists?

2 .  Liter ature R e v ie w a nd E v idence  • 15

 A NSW ER S Box 2.1 EVIDENCE CATEGORIES

1. ANSWER: B Therapy/​Prevention or Etiology/​Harm

Prognosis
Many clinical trials have more than one measured outcome Diagnosis
variable and several demographic variables of interest. Thus,
a number of statistical comparisons will need to be made to Differential Diagnosis or Symptom Prevalence Study
analyze and interpret all of the data. The issue of multiple Economic and Decision Analysis
comparisons arises when enough significance tests are done,
which leads to the increased likelihood that a test will be
statistically significant based on chance alone. Multiple for therapy studies, systematic reviews (with homogeneity)
comparisons include repeated analyses of the same outcome of RCTs receive the highest rating of 1a. Cohort studies and
variable and comparisons of multiple variables, including outcomes research generally receive a rating in the 2 cate-
testing for differences in baseline characteristics and sub- gory, whereas case–​control studies are classified in the 3 cat-
group analyses. The significance level is also known as the egory. A case series would be rated as 4, whereas an expert
type I error rate and is the probability of a false positive. It is opinion paper would receive the lowest rating of 5.
denoted by the Greek letter α.
The implication of multiple comparisons is that the inves-
tigator should be cautious when interpreting the results. One F U RT H E R R E A DI NG
way to counter the problem is to require a lower significance
level; however, this will reduce the power of the trial. Another Centre for Evidenced-​Based Medicine (CEBM). Oxford Centre for
alternative is to increase the sample size so that a smaller sig- Evidence-​Based Medicine—​Levels of evidence (March 2009). 2009.
nificance level can be used while maintaining the power of Available at [Link] [Link]/​oxford-​centre-​evidence-​based-​
medicine-​levels-​evidence-​march-​2009. Accessed July 21, 2015.
the trial. This option may prove to be quite difficult for most
investigators. Many adjustments can be used to approximate
or control the significance level that should be used for inter-
pretation of significant findings, including the Bonferroni 3. ANSWER: D
correction, Holm procedure, and Hochberg procedure.
In the case of this investigator who is running 100 sepa- Randomized control trials are comparative studies with
rate comparisons with a significance level of 0.05, 5 of them an intervention group and a control group, in which the
will be significant based on chance alone. assignment of a subject to a group is determined by a formal
process of randomization. In the simplest of terms, random-
ization is a procedure in which all participants are equally
F U RT H E R R E A DI NG likely to be assigned to either the intervention group or the
control group.
Cook TD, DeMets DL. Selected issues in the analysis. In: Introduction to Randomization is an important concept and is advan-
Statistical Methods for Clinical Trials. Boca Raton, FL: Chapman &
Hall/​CRC Press; 2008: 333–​370.
tageous for many reasons. First, randomization tends to
Friedman LM, Furberg CD, DeMets DL. Issues in data analysis. produce comparable groups. This means that measured
In: Fundamentals of Clinical Trials. 4th ed. New York, NY: Springer; and unmeasured or unknown characteristics and prognos-
2010:345–​390. tic factors of the participants will be, on average, evenly
balanced between the intervention and control groups.
Second, randomization removes the possibility of bias in
2. ANSWER: B the allocation of participants to the intervention group or
to the control group, also known as selection bias. Selection
Utilizing a common framework for judging the strength of bias can be conscious or subconscious and can easily invali-
scientific work allows for a standard approach to compar- date comparisons, which is why randomization is so impor-
ing evidence and allows conflicting evidence to be weighted tant. Finally, randomization provides a sound foundation
differentially. This systematic approach to levels of evidence for valid statistical inference and guarantees the validity of
is essential to the practice of evidence-​based medicine. The inferential tests of statistical significance. Thus, it ensures
Oxford Centre for Evidence-​Based Medicine publishes and independence between assigned treatment and outcome
maintains a “Levels of Evidence” document. and allows the researcher to state that observed differences
Evidence is classified into one of five categories (Box 2.1). between treatment groups are not attributable to chance.
The evidence is then rated, in declining order of strength, as Many different procedures can be employed to provide
1a, 1b, 1c, 2a, 2b, 2c, 3a, 3b, 4, or 5. The ratings are based on randomization for research studies, including blocked,
the design and quality of the study or paper. For example, stratified, adaptive, and play the winner.

16 • Pa in M edicine B oa rd R e v ie w
 F U RT H E R R E A DI NG information between the investigator and the subject.
Institutional review boards (IRBs), clinical investigators,
Cook TD, DeMets DL. Randomization. In: Introduction to Statistical and research sponsors all share responsibility to ensure that
Methods for Clinical Trials. Boca Raton, FL: Taylor & Francis;
2008:141–​170.
the informed consent process is adequate. The US Food and
Friedman LM, Furberg CD, DeMets DL. Issues in data analysis. Drug Administration’s Code of Federal Regulations Title
In: Fundamentals of Clinical Trials. 4th ed. New York, NY: Springer; 21 provides the required basic and optional additional ele-
2010:345–​390. ments of the informed consent document (Box 2.2).
Piantadosi S. Treatment allocation. In: Clinical Trials: A Methodologic
Perspective. 2nd ed. Hoboken, NJ: Wiley; 2005:331–​353.

Box 2.2 BASIC AND ADDITIONAL ELEMENTS

OF INFOR MED CONSENT

4. ANSWER: A • Basic elements

• A statement that the study involves research, including
Although randomized, placebo-​controlled trials are highly • Explanation of the purposes of the research
regarded to be the gold standard of clinical research, the • Expected duration of the subject’s participation
use of a placebo in surgery and skill-​dependent therapies, • Description of the procedures to be followed
such as interventional pain management, is considered to • Identification of experimental procedures
be controversial due to the ethical considerations of per- • Description of reasonable foreseeable risks or
forming a “sham” intervention. Sham interventions are apt discomforts to the subject
to cause moral discomfort in clinician–​investigators, who • Description of any benefits to the subject or others that
are trained to perform invasive interventions only for the may be expected from the research
medical benefit of patients. • Disclosure of alternative procedures or treatment that
Although sham interventions have the potential to harm may be advantageous to the subject
subjects, research designs without a placebo or sham interven- • A statement describing how confidentiality of the
tion are considered to be scientifically less vigorous. Horng and records identifying the subject will be maintained
Miller contend that ethical objections—​based on risk–​benefit • For research involving more than minimal risk
optimization and informed consent issues—​do not support • Explanation as to whether compensation and any
an absolute prohibition of the use of placebo/​sham interven- medical treatments are available if injury occurs and
tions when their use is methodologically necessary to answer • If so, what they consist of and where further
clinically relevant questions. They suggest that each proposed information can be obtained
trial involving sham procedures must be carefully evaluated in • Contact information for answers to pertinent questions
light of these ethical considerations. Furthermore, Miller and about the research and research subject’s rights and
Kaptchuk purport that the use of sham interventions does whom to contact in the event of a research-​related injury
not violate the rights of patient–​subjects provided they have • A statement that participation is voluntary
been adequately informed and fully understand that they will • Refusal to participate will involve no penalty or loss
receive either a real or a sham intervention. of benefits to which the subject is otherwise entitled
• Subject may discontinue participation at any time
without penalty
• Additional elements
F U RT H E R R E A DI NG • A statement that there may be unforeseeable risks to

Horng S, Miller FG. Is placebo surgery unethical? N Engl J Med. the subject
2002;347:137–​139. • Circumstances in which the subject’s participation may
Miller FG, Kaptchuk TJ. Sham procedures and the ethics of clinical be terminated by the investigator
trials. J R Soc Med. 2004;97(12):576–​578. • Any additional costs to the subject that may result from
Piantadosi S. Contexts for clinical trials. In: Clinical Trials:
A Methodologic Perspective. 2nd ed. Hoboken, NJ: Wiley; participation in the research
2005:65–​105. • Consequences of a subject’s decision to withdraw from
the research
• Procedures for orderly termination of participation
by the subject
5. ANSWER: C
• A statement that significant new findings developed
during the course of research that may relate to their
Informed consent is an essential and important facet of
willingness to continue participation will be provided
clinical research. Instead of being viewed as a simple end-
• The approximate number of subjects involved in
point of a signature on a form, the consent document
the study
should be viewed as a basis for meaningful exchange of

2 .  Liter ature R e v ie w a nd E v idence  • 17

 F U RT H E R R E A DI NG other outcomes in clinical trials. In order to relieve clini-
cal symptoms, the objectives of health care interventions
US Department of Health and Human Services, Office for Human include improvement of functioning and health-​related
Research Protections. Informed consent checklist. Available at
[Link] [Link]/​ohrp/​policy/​[Link]. Accessed July
quality of life. IMMPACT was formed with the mission
10, 2015. to develop consensus reviews and recommendations for
US Food and Drug Administration. A guide to informed improving the design, execution, and interpretation of clin-
consent—​ Information sheet. Available at [Link] [Link]/​ ical trials for treatments of pain. IMMPACT recommenda-
RegulatoryInformation/​Guidances/​[Link]. Accessed
July 10, 2015.
tions and guidelines have been widely cited and have helped
US Food and Drug Administration. CFR—​Code of Federal Regulations guide clinical trial design. Specific areas in which they have
Title 21. Available at [Link] [Link]/​scripts/​cdrh/​ made recommendations include core outcome domains,
cfdocs/​cfcfr/​[Link]?fr=50.25. Accessed July 10, 2015. core outcome measures, development of outcome measures,
interpretation of clinical importance of treatment out-
comes, core outcome and treatment measures for pediatric
6. ANSWER: B pain, clinical importance of group differences, analyzing
multiple endpoints, research design for confirmatory clini-
When making an evidence-​based recommendation, it is cal trials, research design for proof-​of-​concept studies, and
important to be able to summarize the quality of the under- design implications for chronic pain prevention studies.
lying evidence. The Oxford Centre for Evidence-​Based With regard to the physical functioning domain of
Medicine has designed a system for grading recommenda- health-​related quality of life, measures include the ability
tions. Grades of recommendation range, in declining order to carry out such daily activities as household chores, walk-
of strength of underlying evidence, from A to D. The grad- ing, work, travel, and self-​care, in addition to strength and
ing takes into account the levels of evidence assigned to endurance measures. Based on the 2005 IMMPACT con-
the underlying studies. Also taken into account is whether sensus publication, the group recommends use of either
extrapolation has occurred or whether the intended use sit- the Multidimensional Pain Inventory or the Brief Pain
uation has potential clinically important differences from Inventory (BPI) for physical functioning measures. The
the original study situation. This type of extrapolation gen- BPI contains a Pain Interference Scale that provides reliable
erally results in a one category downgrade of the grade of and valid measures of the interference of pain with physi-
recommendation. The Oxford Centre for Evidence-​Based cal functioning. The SF-​36 Health Survey may be used as
Medicine has summarized this grading scale (Table 2.1). a more generic measure of health-​related quality of life per
IMMPACT’s guidelines.
Table 2 .1 GR ADES OF R ECOMMENDATION

GR ADE DESCR IPTION

F U RT H E R R E A DI NG
A Consistent level 1 studies
B Consistent level 2 or 3 studies or extrapolations from Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for
level 1 studies chronic pain clinical trials: IMMPACT recommendations. Pain.
2005;113:9–​19.
C Level 4 studies or extrapolations from level 2 or 3 Initiative on Methods, Measurement, and Pain Assessment in Clinical
studies Trials (IMMPACT). IMMPACT website. [Link] [Link].
org/​[Link]. Accessed July 13, 2015.
D Level 5 evidence or troublingly inconsistent/​
inconclusive studies of any level

8. ANSWER: C
F U RT H E R R E A DI NG
Treatment response is often highly variable among subjects
Centre for Evidenced-​Based Medicine (CEBM). Oxford Centre for in clinical trials for pain. Variability may be due to multiple
Evidence-​Based Medicine—​Levels of evidence (March 2009). 2009.
Available at [Link] [Link]/​oxford-​centre-​evidence-​based-​ factors, including different degrees of improvement due to
medicine-​levels-​evidence-​march-​2009. Accessed July 21, 2015. placebo effects or other nonspecific factors, protocol adher-
ence, difficulty in reliable and consistent pain reporting,
difficulty tolerating the treatment, and treatments work-
ing better in some individuals than in others. The enrich-
7. ANSWER: E ment study design can be used in an attempt to decrease
these various sources of variability in order to increase the
In a clinical trial for chronic pain, pain reduction is a neces- chances of detecting an effect if it truly exists. An enrich-
sary outcome variable; however, it is important to consider ment design uses run-​in periods to identify and exclude

18 • Pa in M edicine B oa rd R e v ie w
subjects who have a prespecified level of treatment or pla- possible that the individuals who are willing to volunteer
cebo response, noncompliance, treatment intolerability, for research are different in characteristics than the popula-
or variability in pain ratings. Thus, the enrichment design tion as a whole. For example, it may be that more impov-
helps a researcher select subjects in whom a treatment effect erished individuals volunteer for a paid drug trial, whereas
may be more easily detected. affluent individuals do not.
In the case of this study, by enriching the study with
subjects based on a prespecified level of positive response
to the investigational treatment, the second phase of the F U RT H E R R E A DI NG
study will consist of a cohort of subjects for whom the treat-
ment is likely to be efficacious. Limitations of enrichment Centre for Evidence-​Based Medicine (CEBM). Study designs. 2014.
study design include a limit on the generalizability of the Available at [Link] [Link]/​study-​designs. Accessed July
results to a larger population of patients and the possibility 21, 2015.
University of Texas at Austin. Common mistakes in using statis-
of unblinding in the second phase as a patient who responds tics: Biased sampling. 2015. Available at [Link] [Link].
to a certain treatment may recognize the absence of pain edu/​users/​mks/​statmistakes/​[Link]. Accessed July
relief or side effects if he or she is switched to placebo. 21, 2015.

F U RT H E R R E A DI NG
10. ANSWER: D
Dworkin RH, Turk DC, Peirce-​Sandner S, et al. Research design con-
siderations for confirmatory chronic pain clinical trials: IMMPACT The main purpose of clinical research is to perform a study
recommendations. Pain. 2010;149:177–​193.
Gewandter JS, Dworkin RH, Turk DC, et al. Research designs for in which the results obtained are applicable to a target popu-
proof-​of-​concept chronic pain clinical trials: IMMPACT recom- lation of people with a certain disease. Practically speaking,
mendations. Pain. 2014;155:1683–​1695. it is usually not possible to perform a study on “all” people
in the target population. Instead, studies are performed on
a sample of people drawn from the target population. The
results of a clinical study are therefore used as estimates
9. ANSWER: D of what may happen if the treatment is given to the whole
population of interest. Confidence intervals (CIs) provide a
Randomized controlled trials are experimental studies in range of plausible values for a population parameter based
which subjects are randomly assigned to treatment/​inter- on the study data results and give an indication of the pre-
vention groups or control/​placebo groups. RCTs are among cision of the measured treatment. The 95% CI is usually
the most rigorous study designs that can be employed. The reported in the medical literature and represents the range
advantages of such a design include increased likelihood of in which there is 95% certainty that the true population
blinding (particularly in double-​blinded studies), statistical parameter will lie. The width of a CI indicates the preci-
analysis facilitated by randomization, and unbiased distri- sion of the estimated parameter in that the wider the CI,
bution of confounders. the less the precision and higher amount of random error in
The disadvantages include increased time to conduct the the measurements.
research; increased expense; and occasionally ethical rami- CIs also provide useful information on the clinical
fications having to do with randomizing patients to differ- importance of the results and, like p values, can be used
ent treatment or nontreatment/​placebo groups, especially if to assess statistical significance. Clinical significance is
there is concern that one treatment option may be clearly represented by a difference in effect size between groups
superior. Attrition bias may occur when patients drop out that could be considered important in clinical decision-​
of the study from one or the other of the study groups pref- making. If an effect size is known as being clinical impor-
erentially. For example, participants in the control group tant, CIs that contain that effect size values can indicate
may be unhappy with a lack of progress and may drop out that the result of the test is likely of clinical significance.
of the study to seek alternative treatment or participants in p values provide no information on the clinical impor-
the treatment group may become lost to follow-​up if treat- tance of any observed differences between study groups.
ment has been successful. Whereas a p value indicates whether the results could or
An additional disadvantage is volunteer bias. The goal could not have arisen by chance, a statistically significant
of sampling is to obtain a representative sample of the larger finding has little relation to clinical significance. CIs can
population to be studied. Randomization is employed to provide information on statistical significance in that a
improve the quality of the sample. However, patients are p value will be less than 0.05 if the CI does not include
being randomized from a sample population consisting of whatever value is specified in the null hypothesis. For
volunteers willing to participate in the research. It may be example, if a CI for a mean difference does not include 0,

2 .  Liter ature R e v ie w a nd E v idence  • 19

the data are not consistent with equal population means, Turk DC, Dworkin RH, Allen RR, et al. Core outcome domains for
and we can state that there is a statistically significant dif- chronic pain clinical trials: IMMPACT recommendations. Pain.
2003;106(3):337–​3 45.
ference between the groups.

F U RT H E R R E A DI NG
12. ANSWER: D
Akobeng AK. Confidence intervals and p-​values in clinical decision
making. Acta Paediatr. 2008;97:1004–​1007. Parametric and nonparametric are two broad classifica-
Gardner MJ, Altman DG. Confidence intervals rather than P val- tions of statistical testing procedures. Parametric tests
ues: Estimation rather than hypothesis testing. Br Med J (Clin Res are based on assumptions about the distribution of the
Ed). 1986;292:746–​750.
underlying population from which the sample was taken.
The most common assumption is that the data are nor-
mally distributed, also known as a Gaussian distribu-
tion. Characteristics of normal distributions include
11. ANSWER: E
the following : Data are symmetric about the mean,
have bell-​shaped density curves with a single peak, and
IMMPACT was formed with the mission to develop con-
are defined by mean (μ) and standard deviation (σ); and
sensus reviews and recommendations for improving the
mean, median, and mode are the same. In normal distri-
design, execution, and interpretation of clinical trials for
butions, 68% of the total area under the curve is within
treatments of pain. IMMPACT recommendations and
one standard deviation of the mean, 95% of the total
guidelines have been widely cited and have helped guide
area under the curve is within two standard deviations
chronic pain clinical trial design. Specific areas in which it
of the mean, and 99.7% of the total area under the curve
has made recommendations include core outcome domains,
is within three standard deviations of the mean. Other
core outcome measures, development of outcome measures,
factors that determine whether or not a parametric test
interpretation of clinical importance of treatment out-
is suitable include the type of data being analyzed, homo-
comes, core outcome and treatment measures for pediatric
geneity of variances, and whether or not the samples are
pain, clinical importance of group differences, analyzing
independent. In contrast, nonparametric statistical pro-
multiple endpoints, research design for confirmatory clini-
cedures rely on few or no assumptions about the shape
cal trials, research design for proof-​of-​concept studies, and
or parameters of the population distribution from which
design implications for chronic pain prevention studies.
the sample was taken.
Turk et al. recommended that each of the six core out-
It is important to understand when to use a para-
come domains should be considered in all clinical trial
metric versus a nonparametric statistical procedure.
designs for both efficacy and effectiveness of treatments for
Nonparametric tests use less information and therefore are
chronic pain. Furthermore, if one or more of the domains
more conservative tests compared to their parametric alter-
are not used as an outcome in a study, the reasons for exclud-
natives. Thus, if a nonparametric test is used when you have
ing the outcome should be justified a priori. The six core
parametric data, the power of the analysis can be decreased,
outcome domains as recommended by IMMPACT are
meaning you are less likely to get a significant result when
pain, physical functioning, emotional functioning, partici-
there truly is a significant result. However, if a parametric
pant ratings of global improvement and satisfaction with
test is used wrongly when the data are actually nonpara-
treatment, symptoms and adverse events, and participant
metric, the likelihood of incorrect conclusions increases. In
disposition. Additional or supplemental outcome domains
addition to less power, results of nonparametric procedures
that researchers may elect to use include role functioning,
are more difficult to interpret because many of the tests
interpersonal functioning, pharmacoeconomic measures
use rankings of the values in the data rather than using the
and health care utilization, biological markers, coping, cli-
actual data, which reduces the clinical understanding of the
nician ratings of global improvement, neuropsychological
data and results.
assessments of cognitive and motor function, and suffering
or other end-​of-​life issues.

F U RT H E R R E A DI NG
F U RT H E R R E A DI NG
Hoskin T. Parametric and nonparametric: Demystifying the terms.
Initiative on Methods, Measurement, and Pain Assessment in Clinical Available at [Link] [Link]/​mayo-​edu-​docs/​center-​for-​
Trials. IMMPACT website. [Link] [Link]/​[Link]. translational-​science-​activities-​documents/​berd-​5-​[Link]. Accessed
Accessed July 13, 2015. July 23, 2015.

2 0 • Pa in M edicine B oa rd R e v ie w
13. ANSWERS: E. The prevalence of a disease or risk factor can be
quantified best with cross-​sectional survey. See
A. A controlled trial in which each subject has both “cross-​sectional survey” row in Table 2.2.
therapies at various points in time is a crossover F. The only feasible study design for the study of very
design. Studies with a crossover design allow rare disorders is a case–​control study. See “case–​
each subject to receive both therapies. They are control study” row in Table 2.2.
randomized to treatment A or treatment B first and
then switch to the other treatment at the crossover
point. Subjects serve as their own controls, and all
F U RT H E R R E A DI NG
subjects receive treatment at least part of the time.
This design can be problematic if the washout period Centre for Evidence-​Based Medicine (CEBM). Study designs. 2014.
for a treatment is lengthy or unknown. In addition, Available at [Link] [Link]/​study-​designs. Accessed July
the treatment effect has to be reversible. In other 21, 2015.
words, if the treatment could lead to a permanent
cure for a condition, then a crossover study design is
not appropriate.
B. The design that is best for studying the effect of an 14. ANSWER: B
intervention is the randomized controlled trial. See
“randomized controlled trial” row in Table 2.2. In this study, the analysis to be performed is to compare
C. The study design in which data are obtained from means between two independent groups. Given that the
groups that have been exposed or not exposed to data are normally distributed, one can utilize a parametric
a variable of interest is a cohort study. See “cohort test to compare the two groups. The appropriate parametric
study” row in Table 2.2. statistical procedure to compare the means of two indepen-
D. The study design that is best for the study of the dent groups is a Student’s t-​test. In Table 2.3, common anal-
effect of predictive risk factors on an outcome is a ysis types and statistical procedures are categorized with
cohort study. See “cohort study” row in Table 2.2. corresponding parametric and nonparametric tests.

Table 2 .2 ATTRIBUTES OF R ESEARCH STUDY DESIGNS

STUDY DESIGN DESCR IPTION ADVANTAGES DISADVANTAGES

Randomized An experimental comparison study in • Unbiased distribution of • Time-​consuming and costly

controlled which randomization is used to allocate confounders • Volunteer bias
trial(RCT) participants into control/​placebo or • Blinding
intervention/​treatment groups. This is • Statistical analysis facilitated
the best design to study the effect of an by randomization
intervention.
Crossover Each study participant has both • Subjects serve as their own • Not a good design for nonreversible
design treatments or interventions. Subjects control, leading to reduced outcomes where treatment or
are randomized to which treatment they sample size requirements. intervention leads to a permanent
receive first. At the crossover point, the • All subjects receive the change or cure
subjects switch treatments. treatment or intervention at • Not good for treatments with
some point. lengthy or unknown washout
• Blinding period
Cohort study Subjects are identified who have already • Timing and directionality of • Difficult to identify controls
been exposed, or not exposed, to the events can be established. and blind
factor. Data are then collected. The • It is more convenient and less • Possibility of hidden confounder
design is best for determining the effect expensive than an RCT. • No randomization
of predictive risk factors on an outcome. • Not suited for rare diseases
Case–​control A careful process is used to identify • Fast and inexpensive • Relies on historical data or recall to
study patients with or without a particular • Feasible for rare disorders determine exposure
disease or outcome. Data are then • Smaller sample size required in • Possibility of confounders
collected on exposure to factors of comparison to cross-​sectional • Potential recall and selection bias
interest. study
Cross-​sectional Exposure and outcome are measured at • Inexpensive • Established association not
survey the same time in a defined population. • Simple causality
This is best for determining the • No ethical implications • Possible recall bias
prevalence of a disease or risk factor. • Neyman bias

2 .  Liter ature R e v ie w a nd E v idence  • 21

Table 2 .3 PAR AMETRIC AND NONPAR AMETRIC STATISTICAL PROCEDUR ES BY TYPE OF ANALYSIS

ANALYSIS TYPE EX AMPLE PAR AMETR IC TEST NONPAR AMETR IC TEST

Compare means between two Is the mean pain score at baseline for patients Student’s t-​test Wilcoxon rank-​sum test
independent groups assigned to treatment group different from
the mean pain score for patients assigned to
the placebo group?
Compare two numerical Was there a significant change in quality of Paired t-​test Wilcoxon signed-​rank test
measurements taken from life scores between baseline and the 3-​month
the same individuals follow-​up measurement in the treatment
group?
Compare means between three Do baseline physical functioning scores differ at Analysis of variance Kruskal–​Wallis test
or more independent groups baseline in an experiment with three distinct (ANOVA)
groups (placebo, drug 1, and drug 2)?
Compare multiple numerical Was there a significant change in pain scores Repeated measures Friedman test
measurements taken from the in patients receiving treatment measured at ANOVA
same individuals baseline, 1 month, 3 months, and 6 months?

F U RT H E R R E A DI NG Turner L, Shamseer L, Altman DG, et al. Consolidated Standards of

Reporting Trials (CONSORT) and the completeness of reporting
Hoskin T. Parametric and nonparametric: Demystifying the terms. of randomised controlled trials (RCTs) published in medical jour-
Available at [Link] [Link]/​mayo-​edu-​docs/​center-​for-​ nals. Cochrane Database Syst Rev. 2012; Issue 11:1–​162.
translational-​science-​activities-​documents/​berd-​5 -​6 .pdf. Accessed
July 23, 2015.
Sheskin DJ. Handbook of Parametric and Nonparametric Statistical
16. ANSWER: B
Procedures. 5th ed. New York, NY: Chapman & Hall/​ CRC
Press; 2011. In determining clinically important changes for outcome
measures in the study of pain, interpretation of two sepa-
rate aspects of the results must be distinguished. First, it
15. ANSWER: C must be established what change in the outcome measure
represents a clinically meaningful difference for patients.
Comprehension of the results of an RCT entails that readers Second, it must be established what difference in the mag-
must have a complete understanding of its design, conduct, nitude of the response between the control and treatment
analysis, and interpretation. In order for this to occur, the groups is deemed to be large enough to ascertain the thera-
authors of the study must provide complete transparency of peutic significance of the results.
all details. In the mid-​1990s, CONSORT was developed by For clinical trials designed to evaluate the efficacy of
an international group of clinical trialists, statisticians, epi- chronic pain therapies, the primary outcome of inter-
demiologists, and biomedical journal editors as a means of est commonly involves reduction in pain score intensity.
improving reporting of RCTs. The CONSORT Statement, Multiple studies have been performed to evaluate the mag-
most recently updated in 2010, is an evidence-​based mini- nitude of pain reduction that represents a clinically mean-
mum set of recommendations including a checklist and ingful response to the pain treatment for patients, and an
flow diagram for reporting RCTs (Figure 2.1). It is meant IMMPACT publication summarized the results with rec-
to facilitate the complete and transparent reporting of trials ommendations as noted in Box 2.3.
and aid in their critical appraisal and interpretation.
Although use of the CONSORT Statement is not uni-
versal and not required by all medical journals and texts, Box 2.3 CLINICALLY MEANINGFUL CHANGES
a Cochrane review by Turner et al. concluded that jour- IN VISUAL ANALOGUE SCALE PAIN SCOR ES
nal endorsement of the CONSORT Statement may ben- • Minimal clinically meaningful change
eficially influence the comprehensiveness of RCT and trial • Raw pain score change of approximately 1 point
reporting published in medical journals and texts. • 10–​20% reduction in chronic pain intensity
• Moderate clinically meaningful change
• Raw pain score change of approximately 2 points

F U RT H E R R E A DI NG • ≥30% reduction in chronic pain intensity

• Substantial clinically meaningful change
Moher D, Schulz KF, Altman D. The CONSORT statement: Revised • Raw pain score change of approximately 4 points
recommendations for improving the quality of reports of parallel-​ • ≥50% reduction in chronic pain intensity
group randomized trials. JAMA. 2001;285(15):1987–​1991.

2 2 • Pa in M edicine B oa rd R e v ie w
 CONSORT 2010 Flow Diagram

Enrollment
Assessed for eligibility (n = )

Excluded (n = )
Not meeting inclusion criteria (n = )
Declined to participate (n = )
Other reasons (n = )

Randomized (n = )

Allocation
Allocated to intervention (n = ) Allocated to intervention (n = )
Received allocated intervention (n = ) Received allocated intervention (n = )
Did not receive allocated intervention (give Did not receive allocated intervention (give
reasons) (n = ) reasons) (n =)

Follow-Up

Lost to follow-up (give reasons) (n = ) Lost to follow-up (give reasons) (n = )

Discontinued intervention (give reasons) (n = ) Discontinued intervention (give reasons) (n = )

Analysis

Analysed (n = ) Analysed (n = )
Excluded from analysis (give reasons) (n = ) Excluded from analysis (give reasons) (n = )

Figure 2.1 The CONSORT flow diagram.

SOU RCE: CONSORT Group ([Link] [Link]-​[Link]/​consort-​statement/​flow-​d iagram).

17. ANSWER: A
F U RT H E R R E A DI NG
The matter of which participants are to be included in a
Dworkin RH, Turk DC, Wyrwich KW, et al. Interpreting the study’s data analysis often arises during clinical research tri-
clinical importance of treatment outcomes in chronic pain
clinical trials: IMMPACT recommendations. J Pain. 2008;9(2): als. Even the most carefully managed and well-​designed trial
105–​121. cannot be perfectly executed. The protocol may not be exactly
Farrar JT, Young JP Jr, LaMoreaux L, et al. Clinical Importance of adhered to, outcome and response variable data may be miss-
changes in chronic pain intensity measured on an 11-​point numeri- ing, and some patients may not actually have been eligible for
cal pain rating scale. Pain. 2001;94:149–​158.
Hanley MA, Jensen MP, Ehde DM, et al. Clinically significant changes the study based on inclusion and exclusion criteria issues. This
in pain intensity ratings in persons with spinal cord injury or ampu- can lead to controversy when planning the statistical analy-
tation. Clin J Pain. 2006;22:25–​31. ses for the data because these problems can introduce bias and
Salaffi F, Stancati A, Silvestri CA, et al. Minimal clinically potentially disrupt the validity of the results.
important changes in chronic musculoskeletal pain inten-
sity measured on a numerical rating scale. Eur J Pain. 2004;8: The intention-​to-​treat principle states that all partici-
283–​291. pants randomized and all events should be accounted for in

2 .  Liter ature R e v ie w a nd E v idence  • 2 3

the primary analysis based on the group to which participants F U RT H E R R E A DI NG
were randomized, regardless of whether or not they adhered
to the assigned intervention. Following the intention-​to-​treat Laerd Statistics. Understanding the different types of variable in sta-
tistics. 2015. Available at [Link]
principle may underestimate the full effect of the treatment, guides/​t ypes-​of-​[Link]. Accessed July 27, 2015.
but it guards against the more pressing issue of biased results.
A per protocol analysis only includes participants who were
fully adherent to the protocol with regard to the assigned
study medication, follow-​up visits and/​or measurements, 19. ANSWER: C
and had no other protocol violations. The main issue with
per protocol analysis is that participants who adhere to study In a double-​blinded study, both patients and providers
treatment and protocol may be different than those who are unaware of the patients’ group assignment. In a triple-​
drop out in ways that are related to the outcome of interest. If blinded study, another group, such as support staff or data
the results of intention-​to-​treat and per protocol analyses are analyzers, is also blinded. Randomization is a process of
different, then the intention-​to-​treat analysis results typically selecting from a group in a manner that makes equal dis-
predominate for estimates of efficacy because they maintain tribution of confounders likely. Randomization will not
the value of randomization. Unlike per protocol analysis, work as well with small groups. Stratification is a strategy
intention-​to-​treat analyses can only bias the estimated effect in which patients are intentionally divided by an impor-
in the conservative direction by favoring the null hypothesis. tant characteristic prior to randomization. Confounding
occurs when study results are influenced by a factor other
F U RT H E R R E A DI NG than that which is being studied. When a sample is skewed,
it may be due to sampling error or related bias. For exam-
Friedman LM, Furberg CD, DeMets DL. Issues in data analysis. ple, if diabetic patients are more likely than nondiabetics to
In: Fundamentals of Clinical Trials. 4th ed. New York, NY: Springer;
2010:345–​390.
volunteer for a study, that is a form of volunteer bias.
Grady D, Cummings SR, Hulley SB. Alternative trial designs and
implementation issues. In: Hully SB, Cummings SR, Browner
SR, et al. (Eds.), Designing Clinical Research. 3rd ed. Philadelphia, F U RT H E R R E A DI NG
PA: Lippincott Williams & Wilkins; 2007:163–​182.
Piantadosi S. Counting subjects and events. In: Clinical Trials: A
Stomp on Step 1. Confounding, randomization & blinding. 2015.
Methodologic Perspective. 2nd Ed. Hoboken, NJ: Wiley; 2005:395–​407.
Available at [Link] [Link]/​confounding-​placebo-​
stratification-​randomization-​blinding. Accessed July 27, 2015.

18. ANSWER: C

Variables can be divided broadly into numerical and cate- 20. ANSWER: E
gorical categories with further subclassifications as outlined
in Table 2.4. Four situations are possible when performing statistical tests
to try to reject the null hypothesis in favor of the alternative
Table 2 .4 CLASSIFICATION OF VARIABLES hypothesis when interpreting the results of a study. In two
of these situations, assuming the study is free from bias, the
CLASS SUBCLASS DEFINITION findings in the sample and what is reality in the population
Categorical Nominal Variables that have two or are concordant, and the inference by the investigator(s) will
more categories but lack any be correct. However, in the other two instances, a type I or
intrinsic order type II error has been made, and the inference will not be
Example: Male or female correct.
Ordinal Variables that have two or more Prior to beginning the study, the investigator(s) deter-
categories that can be ordered
or ranked
mines a priori what are the maximum chances he or she will
Example: Strongly agree, accept in making type I or type II errors. The probability of
moderately agree, etc. committing a type I error is also known as α or the signifi-
Numerical Interval Numerical values that can be
cance level. A type I error occurs when the null hypothesis
measured along a continuum is rejected when in reality there actually is no association
Example: Temperature between the predictor and outcome variable (a false-​positive
Ratio Numerical values measured along finding). The probability of a type II error is known as β.
a continuum where zero of that A type II error occurs when there is a failure to reject the null
variable indicates that none of hypothesis when in reality an association does exist between
that variable is present predictor and outcome (a false-​negative finding). Power is
Example: Weight
specified as 1 –​β and is the probability of correctly rejecting

2 4 • Pa in M edicine B oa rd R e v ie w
the null hypothesis in the study sample if the actual effect two-​period design. More complex crossover trial designs
in the population is greater than or equal to the effect size. may be employed in various clinical circumstances.
In an ideal world, both α and β would be set at zero, thus The crossover design has multiple advantages for
eliminating the possibility of false-​positive or false-​negative researchers. One advantage is that it minimizes variability
results. In real-​life practice, these values are made as small because each participant serves as his or her own control
as possible, with the caveat that the sample size will need to and the subsequent paired analyses substantially increase
increase as these values decrease. the statistical power of the trial in that fewer participants
are required. Thus, the crossover design takes advantage
of making treatment comparisons based on within-​rather
F U RT H E R R E A DI NG than between-​subject differences. This allows the treatment
difference to be estimated with greater precision and less
Browner WS, Newman TB, Hulley SB. Getting ready to estimate possibility for confounding. Recruitment may also be easier
sample size: Hypotheses and underlying principles. In: Hully with this type of design because all subjects will receive all
SB, Cummings SR, Browner SR, et al. (Eds.), Designing Clinical treatments under investigation, which may be an attractive
Research. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins;
2007:51–​63.
attribute for some patients who are concerned about partic-
Friedman LM, Furberg CD, DeMets DL. Sample size. In: Fundamentals ipating in clinical trials in which they may be randomized
of Clinical Trials. 4th ed. New York, NY: Springer; 2010:133–​167. to a no-​treatment or placebo arm.
Disadvantages of the crossover design are related to the
issue of carryover effects and dropouts. Carryover effects
are the residual influence of the intervention on the out-
21. ANSWER: A come during the period after which it has been discontin-
ued. To reduce carryover effect, the investigator can use an
The crossover design is a special type of RCT in which untreated “washout” period between treatments with the
each study treatment is administered at different times to hope that the outcome variable will return to its baseline
every subject enrolled in the study. Participants in this type before starting the next intervention. Another concern
of design “cross over” or “switch” from one treatment to for carryover effects is if they lead to a permanent change
another by this strategy, with the intent to estimate differ- or cure in the underlying condition of the patient. In this
ences between them. Typically, half of the participants are instance, the treatment during the second period could
randomly assigned to start with one treatment (or control) appear falsely or artificially superior. Finally, the patients’
and then switch to the other treatment (or control). In the condition could change in the second treatment phase of
case presented in this question, half would start with drug the study, which possibly may affect how they respond to
A and switch to drug B, and the other half would start with the second treatment.
drug B and switch to drug A (Figure 2.2). This is the sim- The issue of dropouts is of concern for two reasons.
plest type of crossover trial and is called the two-​treatment, First, the participant is exposed to more drugs or treatments

Treatment with Treatment with

Washout Period
Drug A Drug B

Sample Cohort of Interest

Treatment with Treatment with

Washout Period
Drug B Drug A

Baseline Outcome Outcome Outcome

Measurements Measurements Measurements Measurements

Figure 2.2 Schematic of a two-​treatment, two-​period, crossover randomized trial design.

2 .  Liter ature R e v ie w a nd E v idence  • 25

in a crossover trial, increasing the chance of side effects that hypothesis, the appropriate statistical test based on type
could contribute to dropping out. Second, the study is usu- of predictor and outcome variable in the hypotheses, a
ally longer than a regular RCT, thus providing a longer reasonable effect size between the two study groups, vari-
period of opportunity to drop out. The consequences of ability, and an a priori determination of α and β. Even if
dropouts are more impactful in a crossover design because the exact value for one or more of these steps is uncertain
the data loss is more significant; for example, if a participant or unknown, it is important to estimate the sample size
drops out in the second treatment phase, the participant’s prior to starting the study and early in the design phase.
data cannot be analyzed using only the first phase because Many clinical trials are performed that lack the statistical
that participant was acting as his or her own control and power or ability to detect treatment effects of a magni-
determining a treatment effect cannot occur. tude that has some clinical importance. Conversely, some
sample size estimations may assume an unrealistically
large intervention effect, meaning the power for more
F U RT H E R R E A DI NG realistic and smaller effects will be low. Finally, the dan-
ger in studies with low statistical power is that treatments
Friedman LM, Furberg CD, DeMets DL. Basic study design. that could be beneficial are discarded due to not find-
In: Fundamentals of Clinical Trials. 4th ed. New York, NY: Springer;
2010:67–​96. ing statistical significance and may never be investigated
Grady D, Cummings SR, Hulley SB. Alternative trial designs and again. Due to the approximate nature of sample size cal-
implementation issues. In: Hully SB, Cummings SR, Browner culations, investigators should try to balance being realis-
SR, et al. (Eds.), Designing Clinical Research. 3rd ed. Philadelphia, tic with being conservative in estimating the parameters
PA: Lippincott Williams & Wilkins; 2007:163–​182.
Piantadosi S. Crossover designs. In: Clinical Trials: A Methodologic discussed here.
Perspective. 2nd ed. Hoboken, NJ: Wiley; 2005:515–​527.

F U RT H E R R E A DI NG

22. ANSWER: C Browner WS, Newman TB, Hulley SB. Estimating sample size and
power: Applications and examples. In: Hully SB, Cummings SR,
Browner SR, et al. (Eds.), Designing Clinical Research. 3rd ed.
Certain groups of participants are considered to be particu- Philadelphia, PA: Lippincott Williams & Wilkins; 2007:65–​94.
larly vulnerable to undue influence or coercion in a research Friedman LM, Furberg CD, DeMets DL. Sample size. In:
setting. These groups, as outlined in 45 CFR 46, are children, Fundamentals of Clinical Trials. 4th ed. New York, NY: Springer;
2010:133–​167.
wards of the state, prisoners, pregnant women and fetuses,
persons who are mentally disabled or otherwise cognitively
impaired, and economically or educationally disadvantaged
persons. IRBs that review research studies involving all cat-
egories of vulnerable patients must determine that their use 24. ANSWER: B
is adequately justified and that additional safeguards are
implemented to minimize risks unique to each group. In designing a research study, one of the most important
components is creating selection criteria that define the
population to be studied. This is because of the possible
F U RT H E R R E A DI NG effects of prognostic and selection factors on differences
in outcome. The inclusion criteria define the main char-
US Department of Health & Human Services. CFR—​Code of Federal acteristics of the target population that pertain to the
Regulations Title 45, Part 46. Available at [Link] [Link]/​ research question. Inclusion criteria typically include
ohrp/​humansubjects/​g uidance/​[Link]. Accessed July 27, demographic characteristics (age, gender, and ethnicity),
2015.
US Department of Health and Human Services. IRB Guidebook: clinical characteristics (the disease being studied and
Chapter VI Special Classes of Subjects. Available at [Link] its severity), geographic characteristics (patients from
[Link]/​ohrp/​irb/​irb_​[Link]. Accessed July 27, 2015. investigator’s clinic or hospital or patients outside the
investigator’s practice), and time characteristics (study
time frame from start to finish). Ultimately, inclusion
criteria should be as specific as possible, sensible, used
23. ANSWER: E consistently throughout the study, and provide the basis
for understanding to whom the published results and
There are several variations on how sample sizes are esti- conclusions apply.
mated for a study or experiment, but there are common Exclusion criteria, on the other hand, indicate the
features and steps, including the following: stating the subsets and characteristics of the population that might
null hypothesis and either a one-​or two-​sided alternative interfere with follow-​up efforts, the quality of the data, or

2 6 • Pa in M edicine B oa rd R e v ie w
high risk of possible side effects. In clinical trials, exclu- measure of how likely a test will correctly identify a
sions tend to include specific causes of concern for the condition when it is present. Sensitivity is calculated
safety of the participants. As a good overall rule, having as the number of true positives divided by the sum
as few exclusion criteria as possible helps keep recruit- of true positives and false negatives. Specificity, on
ment simple and preserves the number of potential study the other hand, is a measure of the likelihood that a
subjects. person without a disease will have a negative test. It is
calculated as the number of true negatives divided by
the sum of the true negatives and false positives.
F U RT H E R R E A DI NG

Hulley SB, Newman TB, Cummings SR. Choosing the study sub- F U RT H E R R E A DI NG
jects: Specification, sampling, and recruitment. In: Hully SB,
Cummings SR, Browner SR, et al. (Eds.), Designing Clinical Williams M. Sensitivity and specificity: Precision of the clinical exam.
Research. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2015. Available at [Link] [Link]/​EMAC/​curricu-
2007:27–​36. lum/​diagnosis/​[Link]. Accessed July 26, 2015.
Piantadosi S. The study cohort. In: Clinical Trials: A Methodologic
Perspective. 2nd ed. Hoboken, NJ: Wiley; 2005:309–​330.
C. Positive predictive value is calculated as the number
of true positives divided by the sum of true positives
and false positives. It is the probability that a patient
25. ANSWERS: with a positive test actually has the disease. Negative
predictive value, on the other hand, is the number of
true negatives divided by the sum of the true negatives
A. The inverse of ARR is equal to NNT). NNT and false negatives.
represents how many people need to be treated or
exposed to an intervention in order for one person
to have an improved outcome. To calculate NNT, F U RT H E R R E A DI NG
ARR must first be determined. ARR is defined as
Williams M. Sensitivity and specificity: Precision of the clinical exam.
the difference between the control event rate and the 2015. Available at [Link] [Link]/​EMAC/​curricu-
experimental event rate. NNT is equal to the inverse lum/​diagnosis/​[Link]. Accessed July 26, 2015.
of ARR.
D. Power quantifies the likelihood of identifying a
significant effect when it exists. It can be calculated
F U RT H E R R E A DI NG as 1 –​ β.

Centre for Evidence-​Based Medicine. Number needed to treat (NNT).

2014. Available at [Link] [Link]/​number-​needed-​to-​treat-​ F U RT H E R R E A DI NG
nnt. Accessed July 26, 2015.
Calkins K. Power and sample size: Applied statistics. 2015. Available at
B. Specificity correlates to the proportion of negatives [Link] [Link]/​~calkins/​math/​edrm611/​[Link].
that are correctly identified as such. Sensitivity is a Accessed July 26, 2015.

2 .  Liter ature R e v ie w a nd E v idence  • 2 7

 3.
PAIN R ESEARCH
PL AC E B O, A N I M A L MODE L S , E T H IC S , A N D E PI DE M IOL O G Y

David A. Edwards

I N T RODUC T ION 2. A 50-​year-​old female is asked to rate her pain after

receiving a placebo pill. She notices the research
Research of pain requires not only that basic ethical stan- observer watching her and smiling; thus, she reports
dards for human studies research be followed but also an improvement in her pain. This phenomenon is
that special consideration be given due to the potential for known as?
human suffering. The use of placebos as controls became
widespread after World War II with the adoption of the A. Response bias
randomized controlled trial. Other study designs have elu- B. Regression to the mean
cidated the risk factors for pain. The prevalence of pain in C. Hawthorne effect
society is a measure of the burden of pain. D. Mesmerism
Experimentation using animals in pain research has pro- E. Placebo response
vided considerable insight into the pathophysiology of pain
in humans. Because the experience of pain is subjective, the 3. Which of the following is true about the placebo
study of pain in animals must be done in a way that limits effect?
potential suffering. Nociception and the response in animals
must be considered to reflect pain, and so ethical principles to A. It is a psychobiological event that can be attributed
limit potential suffering in animals guide research in this area. to the entire treatment experience.
Historical abuses in human experimentation have B. Patient expectations negatively impact placebo
driven the development of ethical guidelines for human response size.
subjects research. The Belmont Report serves as the modern C. Classical conditioning is the primary reason for the
standard of basic ethical principles on which institutional placebo response.
review boards (IRBs) judge human research studies. D. The Yale–​Brown Compulsive Scale is used to rule
out poor candidates in placebo-​controlled trials.
QU E S T IONS
4. Which of the following is a true definition of nocebo?
1. In randomized controlled trials (RCTs), placebo
response size is often difficult to quantify. Which of A. The paucity of effect from placebo
the following is not a confounder of positive placebo B. Negative side effects of a placebo
response in RCTs? C. An active treatment not meant to cause the effect
observed
A. Response bias D. A nonmedication placebo treatment
B. Regression to the mean E. The observed effect of a placebo
C. Natural course of disease
D. An educated patient 5. A practitioner injects lidocaine into muscle trig-
E. Fluctuation in symptoms ger points and describes to the patient the triggering

28