Designing Clinical Research

ISBN: 978-1-9751-7440-8 | 5th_Edition

(Print pagebreak 1)(Print pagebreak 2)

CHAPTER 1

Getting Started: The Anatomy and Physiology of Clinical

Research
Warren S. Browner
Thomas B. Newman
Mark J. Pletcher

This chapter introduces clinical research from two viewpoints, developing themes that run throughout the book. One is the anatomy
of research—the tangible elements of a study plan, such as the research question, design, participants, measurements, sample size
calculation, and so forth. An investigator’s goal is to design these components in a fashion that will make the project feasible and
efficient and that will enhance the validity of the study’s findings.

The other theme is the physiology of research—how it works. Studies are useful to the extent that they yield valid conclusions about
what happened in the study and then about how those conclusions generalize to the larger world. The goal is to minimize any errors
that might threaten those processes.

Separating the two themes is artificial in the same way that the anatomy of the human body doesn’t make much sense without some
understanding of its physiology. But the separation has the advantage of simplifying our thinking about a complex topic.

▪ ANATOMY OF RESEARCH: WHAT IT’S MADE OF

The structure of a research project is set out in its protocol, the written plan of the study. Protocols are well known as devices for
seeking grant funds and Institutional Review Board (IRB) approval, but they also have a vital scientific function: helping the
investigator organize her research in a logical, focused, and efficient way. Table 1.1 outlines the components of a protocol. We
introduce the whole set here, expand on each component in the ensuing chapters of the book, and return to put the completed pieces
together in Chapter 20.

Research Question
The research question 1 is the objective of the study, the uncertainty the investigator wants to resolve. Research questions
often begin with a general concern that must be narrowed down to a concrete, researchable issue. Consider, for example, the
general question of whether caffeine affects cognitive function. Although a good place to start, this question must be focused
further before planning efforts can begin. Often, this involves breaking the question into more specific components and
singling out one or two of these to build the protocol around. Here are a few examples of possible research questions:

Are the effects of caffeine on cognitive function short-term, long-term, or both?

Does chronic caffeine use reduce the risk of developing dementia?

How is caffeine consumption best measured?

Are there harmful effects from consuming caffeine?

Does it matter whether caffeine is consumed in coffee? Tea? Caffeinated soft drinks?

(Print pagebreak 3)

TABLE 1.1 ANATOMY OF RESEARCH: THE STUDY PLAN

DESIGN COMPONENTS PURPOSE

Research question(s) What question(s) will the study address?

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

Background and significance Why are these questions important?

Design
How is the study structured?
Type of study
Participants

Selection criteria Who are they, and how will they be selected?

Sampling design
Variables

Predictor variables
What measurements will be made?
Confounding variables

Outcome variables
Statistical issues

Hypotheses
How large is the study, and how will it be analyzed?
Sample size

Analytic approach

There are, of course, many more potential research questions about this topic; the investigator’s task is to choose one that is
feasible, important, novel, and ethical to study. These attributes have an easy-to-remember acronym—FINE—as discussed
in Chapter 2.

Background and Significance

Developing a meaningful research question requires becoming knowledgeable about the areas to be studied. Often, talking with
experts and doing a thorough literature review will lead the investigator to modify the research question. So the first step in
designing a clinical research project is determining what is—or is not—already known about the topic at hand. That’s followed by
pondering how to add to that current knowledge. This step involves thinking long and hard, and talking with colleagues and
mentors, about the important but as-yet unanswered questions.

These thoughts will eventually be synthesized and summarized in a brief background and significance section that provides the
rationale for the study. In this section, the investigator cites relevant previous research (especially her own and that of her
colleagues), emphasizing the problems with that research and the uncertainties that remain. She also specifies how the findings of
her proposed study will help resolve these uncertainties, by leading to new scientific knowledge, changing clinical practice, or
influencing practice guidelines or health policy.

Design

There is no best design for all research questions. Rather, the investigator must determine the type of design that will work best for
the chosen research question and her available resources. A fundamental decision is whether to apply an intervention and examine
its effects in a clinical trial or simply to make measurements on the study participants in an observational study (Table 1.2).
Among clinical trial options, the randomized blinded trial is usually the most desirable design, but nonrandomized or unblinded
designs may occasionally be all that are feasible for some interventions.

(Print pagebreak 4)

TABLE 1.2 EXAMPLES OF CLINICAL RESEARCH DESIGNS TO FIND OUT WHETHER CAFFEINE
CONSUMPTION REDUCES THE RISK OF DEMENTIA

EPIDEMIOLOGIC DESIGN KEY FEATURE EXAMPLE

Observational Designs

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

The investigator interviews a group of

Cross-sectional study A group examined at one point in time
A group (usually without the outcome)
Cohort study identified at the beginning of the study and function and examines them at follow-up
followed over time
Two groups selected based on the presence group of people with dementia (the
Case-control study
or absence of an outcome
Clinical Trial Design
Two groups created by a random process
Randomized blinded trial
and a blinded intervention
participants about current and past caffeine
intake and correlates results with scores on
a test of cognitive function.
She measures caffeine consumption in a
group of participants with normal cognitive
visits to see if those who consume caffeine
are less likely to develop dementia.
She compares past caffeine intake in a
“cases”) with that in a group with normal
cognitive function (the “controls”).
She randomly assigns participants with
normal cognitive function to receive
caffeine supplements or a placebo that is
identical in appearance and then follows
both treatment groups for several years to
observe the incidence of dementia.

Among observational studies, two common designs are cross-sectional studies, in which observations are made on a single
occasion, and cohort studies, in which observations are made in a group of participants who are followed over time. This latter
group of studies can be further divided into prospective cohort studies—which begin in the present and follow participants into the
future—and retrospective cohort studies—which examine information collected in the past—although these distinctions are not
always hard and fast. A third common observational design is the case-control study, in which the investigator compares a group of
people who have developed the outcome of interest (the cases) with another group that has not (the controls).

Each research question requires a judgment about which design is the most efficient way to get a reliable answer. The randomized
blinded trial is often held up as the best design for establishing causality and determining the efficacy of an intervention, but there
are many situations for which an observational study is a better choice or the only feasible option: It would not be easy to randomly
assign people to stop or start drinking coffee, for example. The relatively low cost of case-control studies and their suitability for
rare outcomes make them attractive for some questions. These issues are discussed in Chapters 8, 9, 10, 11, 12, 13, 14, each dealing
with a particular set of designs.

A typical sequence for studying a topic begins with observational studies of a type often called descriptive. These studies explore
the lay of the land—for example, describing distributions of health-related characteristics and diseases in the population:

What is the frequency of caffeine consumption among adults 70 years of age and older?

What proportion of older adults have abnormal cognitive function?

Studies of medical tests, such as whether serum levels of a new biomarker correlate with cognitive dysfunction, are a special form
of descriptive studies (see Chapter 13).

(Print pagebreak 5)

Descriptive studies are usually followed or accompanied by analytic studies that evaluate associations, usually to permit
inferences about cause-effect relationships:

Is the average amount of caffeine consumed per day associated with performance on tests of cognitive function?

Are people with dementia less likely to have a history of regular caffeine consumption than are controls with normal
cognitive function?

The final step is often a clinical trial to establish the effects of an intervention:

Do older adult volunteers who are randomly assigned to take caffeine supplements have a lower risk of developing
dementia than those who are assigned to take a placebo?

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

Clinical trials usually occur relatively late in a series of research studies about a given question, because they tend to be more
difficult and expensive, may be more likely to expose participants to harm, and usually address a narrowly focused question that
arose from the findings of observational studies.

It is useful to characterize a study in a single sentence that summarizes the design and research question. If the study has two major
phases, the design for each should be mentioned.

This is a cross-sectional study of the association between caffeine consumption and cognitive function in 60- to 74-year-
old people, followed by a prospective cohort study of whether caffeine use is associated with a lower rate of subsequent
cognitive decline.

Some designs do not easily fit into the categories listed previously, and describing them with the right level of detail in a single
sentence can be challenging. It is worth the effort, however: A concise description of the research question and design clarifies the
investigator’s thoughts and is useful for orienting colleagues and consultants.

Study Participants

Two major decisions must be made in choosing the sample of study participants (Chapter 3). The first is to describe the kinds of
people you wish to study, specifying the inclusion and exclusion criteria that will define your study participants. The second
decision concerns how you will recruit an appropriate number of people from an accessible subset of this population to
participate in the study. For example, a case-control study of caffeine use and dementia might begin (after securing IRB approval)
by reviewing electronic health records at your institution to identify patients with a diagnosis of dementia or cognitive dysfunction
and then contacting them and appropriate controls to see if they are interested in participating.

Variables

Another major set of decisions in designing any study concerns the choice of which variables to measure (Chapter 4). A study of
caffeine consumption, for example, might ask about different types of beverages and “stay alert” pills that contain varying levels of
caffeine and include questions about portion size, frequency, and timing of intake.

In an analytic study, the investigator studies the associations among variables to predict outcomes and to draw inferences about
cause and effect. When considering the association between two variables, the one that occurs first—or is more likely on biologic
grounds to be causal—is called the predictor variable; the other is called the outcome variable. 2 Most observational studies have
many predictor variables (eg, age, sex, race/ethnicity, smoking history, caffeine intake) and several outcome variables (eg, cognitive
function, quality of life).

(Print pagebreak 6)

Clinical trials examine the effects of an intervention, which is a special kind of predictor variable that the investigator manipulates,
such as treatment with caffeine capsules (or matching placebos). This design allows her to use randomization to minimize the
influence of confounding variables—other predictors of the outcome, such as smoking or educational level, that also influence
caffeine intake and may confuse the interpretation of the findings. Confounding, a particularly thorny topic, is discussed in detail in
Chapter 10.

Statistical Issues
The investigator must develop plans for estimating the sample size as well as for managing and analyzing the study data. This
generally involves specifying a research hypothesis (Chapter 5), a version of the research question that provides the basis for
testing the statistical significance of the findings:

Hypothesis: Adults 60 to 74 years of age who consume an average of at least two cups of coffee per day (or an
equivalent amount of caffeine) have better cognitive function than those with lower levels of consumption.

The research hypothesis also allows the investigator to calculate the sample size—the number of participants needed to find a
statistically significant difference in outcomes between the study groups with reasonable probability, given that one exists, an
attribute known as power (Chapter 5). Purely descriptive studies (eg, “What proportion of people with normal cognitive
function drink coffee daily?”) do not involve tests of statistical significance and thus do not require a hypothesis; instead, the
number of participants needed to produce acceptably narrow confidence intervals for means, proportions, or other

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

descriptive statistics can be calculated.

▪ PHYSIOLOGY OF RESEARCH: HOW IT WORKS

The goal of clinical research is to improve our understanding of the real world from the findings of the study (see the lower half of F
igure 1.1). First, the findings of the study must be interpreted, a process that depends on internal validity, the degree to which the
study’s results reflect what happened in the study. Next, the investigator makes inferences from the study conclusions to enhance
our understanding of the real world. This process depends on external validity (or generalizability), the fidelity with which those
conclusions apply to people and events outside the study.

When an investigator plans a study, she reverses the process, working from left to right in the upper half of Figure 1.1 with the goal
of maximizing the validity of these inferences at the end of the study. She designs a study plan in which the choice of research
question, participants, and measurements enhances the external validity of the study and facilitates its implementation to maximize
internal validity. In the following sections, we address design and then implementation before turning to the errors that threaten the
validity of these inferences.

▪ FIGURE 1.1 ▪ The basic structure of clinical research. At each step, choices and errors affect the inferences that
can be made from the completed study.

(Print pagebreak 7)

Designing the Study

Consider this simple descriptive research question about caffeine and cognitive function:

What is the average daily caffeine consumption among adults 70 years of age or older with normal cognitive function?

This question cannot be answered with perfect accuracy because it would be impossible to study the entire target population of all
older adults, never mind that methods of measuring caffeine consumption and cognitive function are imperfect. Rather, the
investigator must settle for a related question that can be answered:

What is the self-reported average daily caffeine consumption ascertained from a mailed questionnaire sent to patients
cared for at the investigator’s institution who are at least 70 years of age and who do not have a diagnosis of dementia
or cognitive dysfunction recorded in their electronic medical records?

The transformation from research question to study plan is illustrated in Figure 1.2. One major component of this transformation is
the choice of a sample of participants to represent the population. Because there are almost always practical barriers to studying the
target population as a whole, the group specified in the protocol must usually be a sample of that population. Using the
investigator’s institution to identify potential participants is a compromise that makes the study feasible, but it has the disadvantage
that it may produce a different pattern of caffeine use than in the general population, even if everyone returns the questionnaire.
More important, the absence of a recorded diagnosis of dementia or cognitive dysfunction certainly does not ensure that someone
has normal cognitive function.

The other major component of the transformation is the choice of variables to represent the phenomena of interest. The variables
specified in the study plan are usually proxies for these phenomena. For example, the decision to use a self-report questionnaire to
assess caffeine consumption in the prior month is a fast and inexpensive way to collect information. However, it is unlikely to be
perfectly accurate because people usually cannot remember exactly how much caffeine they drank the previous month and because

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

that month may not reflect their usual pattern of caffeine consumption.

In short, each of the differences in Figure 1.2 between the research question and the study plan represents a choice to make the study
more feasible and efficient. This increase in practicality, however, comes with a cost: the risk that those choices may produce a less
relevant conclusion because the study answered a question that differed somewhat from the research question.

▪ FIGURE 1.2 ▪ Choices and errors made in the design of the study affect its external validity. If the intended
sample and variables do not sufficiently represent the target population and the phenomena of interest, the ability to
draw inferences to improve our understanding of the problem is reduced.

(Print pagebreak 8)

Implementing the Study

The implementation of the study plan affects the fidelity with which the actual study follows that plan. At issue here is the problem
of a wrong answer to the research question because the way the sample was drawn—or the measurements were made—differed
substantially from how they were designed (Figure 1.3).

The actual sample of study participants is almost always different from the intended sample. Plans to study all eligible clinic patients
without dementia or cognitive dysfunction, for example, could be disrupted by incomplete diagnoses in the electronic medical
record, wrong addresses for the mailed questionnaires, and refusals to participate. Those who are reached and agree to participate
may have a different pattern of caffeine consumption than those not reached or not interested. In addition to these problems, the
actual measurements can differ from the intended measurements. If the format of the questionnaire is unclear, participants may get
confused and check the wrong box, or they may simply omit a question by mistake.

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

▪ FIGURE 1.3 ▪ Choices and errors made in the implementation of the study affect its internal validity. If the
actual sample and measurements do not sufficiently represent the intended sample and variables, the ability to draw
valid conclusions from the study’s findings is reduced.

Causal Inference

Many studies aim to establish a cause-effect association—that is, a predictor causes an outcome—in order to identify clinical or
public health interventions that may improve health. For example, if caffeine use causes a decrease in the risk of developing
dementia, then perhaps we should recommend it widely. A special kind of validity problem arises in studies that attempt to establish
causal inference. If a cohort study finds an association between caffeine intake and dementia, does this represent cause and effect, or
is caffeine intake only related to a confounding exposure (say, reading the morning newspaper) that protects against dementia?
Reducing the likelihood of confounding and other rival explanations is one of the major challenges in designing an observational
study (Chapter 10).

The Errors of Research

Recognizing that no study is entirely free of errors, the goal is to maximize the validity of any inferences made from what was
observed in the study. Some erroneous inferences can be addressed in the analysis phase of research, but a better strategy is to focus
on design and implementation (Figure 1.4), preventing errors from occurring in the first place.

(Print pagebreak 9)

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

▪ FIGURE 1.4 ▪ The physiology of research. Choices and errors made in the design and implementation of the study
affect its internal and external validity.

The two main kinds of mistakes that interfere with research inferences are called random error and systematic error. The
distinction is important because the strategies for minimizing them are quite different.

Random error is a mistake due to chance—meaning that there is no known cause or predictable pattern to the error. Random error
can distort a measurement in either direction. For example, if the true prevalence of daily caffeine consumption is exactly 40%
among the several thousand members of the health plan 70 years of age and older, a well-designed sample of 100 patients from that
accessible population might contain exactly 40 patients who consume caffeine each day. It’s quite likely, however, that the sample
would contain a slightly different number, such as 38, 39, 41, or 42. Occasionally, chance would produce a substantially different
number, such as 32 or 49. Among the many techniques for reducing the influence of random error, the simplest is to increase the
sample size. The use of a larger sample diminishes the likelihood of a substantially wrong result by increasing the precision of the
estimate—the degree to which the observed prevalence approximates 40% each time a sample is drawn. But increasing the sample
size also increases the cost of a study; fortunately, there are other ways to reduce random error, including using better instruments to
make measurements (Chapter 4).

Systematic error is a wrong result due to bias—sources of variation that distort the study findings in one direction. An illustration is
the decision (Figure 1.2) to study patients in a local health plan, where the recording of the diagnoses of dementia and cognitive
dysfunction may have been influenced by the plan’s efforts to avoid (or encourage!) overcoding. Increasing the sample size has no
effect on systematic error. The best way to improve an estimate’s accuracy (the degree to which it approximates the true value) is to
reduce the size of potential biases—a topic to which much of this book is devoted. In addition, the investigator can seek additional
information to assess the importance of possible biases, such as using a second sample drawn from another setting.

The examples of random and systematic error in the preceding two paragraphs are components of selection bias, which threatens
inferences from the study participants to the population. Both random and systematic errors can also contribute to measurement
error, threatening the (Print pagebreak 10)
inferences from the study measurements to the phenomena of interest. An illustration of random measurement error is the variation
in responses that occurs when a questionnaire is administered on several occasions. An example of systematic measurement error is
misestimation of the amount of caffeine consumption due to problems in the questionnaire (eg, failing to ask about energy drinks).
Strategies for controlling these errors are presented in Chapters 3 and 4.

TABLE 1.3 APPROACHES TO CONTROLLING RANDOM AND SYSTEMATIC ERRORS IN THE DESIGN AND
IMPLEMENTATION PHASES OF A STUDY

TYPE OF ERROR SOLUTION

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

Random Improve precision of measurements (Chapter 4)

Increase sample size (Chapters 5 and 6)
Systematic Improve accuracy and validity of measurements (Chapter 4)
Choose a better design (Chapters 8, 9, 10, 11, 12, 13, 14)
Random and systematic Quality control (Chapter 18)

As summarized in Table 1.3, getting the best possible answer to the research question is a matter of designing, implementing, and
interpreting the study to minimize the magnitude of potential errors.

▪ TRANSLATIONAL RESEARCH
Translational research refers to studies that translate basic science discoveries into clinical practice or that expand clinical research
findings into broad-scale public health or community-based programs. These studies should be designed in the same way as other
clinical research studies, although the number of investigators is often larger and the process itself more complex and iterative.
Successful translational research requires collaborations among investigators along the spectrum of research, all of whom should be
prepared to revise their plans: It’s not unusual for a treatment that looks promising in laboratory animals to cause adverse effects
when tested in people, or for a screening test that works in an academic clinic to be cumbersome to apply or less accurate in a
community setting. The earlier those collaborations begin, the more likely that the investigators will be able to pose an important
research question, develop a worthwhile study plan, and avoid setbacks that could have been anticipated.

Translating from Laboratory to Clinical Research

A host of tools—including DNA sequencing, gene expression arrays, molecular imaging, proteomics, and metabolomics—have
moved from basic science laboratories into clinical investigation. Compared with ordinary clinical research, successful translation of
tests and treatments discovered in bench research usually requires personal laboratory experience or a collaborator with those skills.
Bench-to-bedside research necessitates a thorough understanding of the underlying basic science. Many clinical researchers believe
that they can master this knowledge easily—just as many laboratory-based scientists believe clinical research does not require
special training and expertise. But the skill sets for basic and clinical research barely overlap, and conceiving good translational
research questions requires expertise and close collaboration in more than one area.

For example, suppose a gene that regulates circadian rhythm in mice has been identified and a clinical investigator with expertise in
sleep disturbances wants to study whether variants in the human homolog of that gene affect sleep patterns in people. To develop a
meaningful research question about the potential effects of that gene, she will need collaborators who are familiar with the biology
of the gene and its protein product(s), as well as the advantages and limitations of the various methods of genotyping. Similarly,
suppose a laboratory-based investigator has found a unique pattern of gene expression in lymph nodes obtained from women (Print
pagebreak 11) with breast cancer. To study whether that pattern is useful prognostically, she should collaborate with investigators
who understand the importance of test-retest reliability, sampling and blinding, and the effects of prior probability of disease on the
applicability and clinical utility of her discovery. Finally, a research team interested in testing a new medication may need scientists
familiar with molecular biology, pharmacokinetics and pharmacodynamics, phase I and II clinical trials, and practice patterns in the
relevant field of medicine.

Translating from Clinical to Population Research

Designing studies to apply findings from clinical research to larger and more diverse populations may also require additional
expertise, such as in identifying high-risk or underserved groups, understanding the difference between screening and diagnosis, and
knowing how to implement changes in health care delivery systems. On a practical level, this kind of research usually needs access
to large groups of participants, such as those enrolled in health plans or the residents of an entire county. Support and advice from a
department chair, the chief of the medical staff at an affiliated hospital, the leader of a managed care organization, or the director of
the local health department may be helpful when planning these studies.

Some investigators take a shortcut when designing this type of translational research, for example, planning to expand on results
seen in their own specialty clinic by enrolling participants from similar practices at other university medical centers, rather than
involving practitioners in the community. This is a bit like translating Aristophanes into modern Greek—it will still not be very
useful for English-speaking readers. Testing research findings in larger and more diverse populations often requires partnerships
with community members and adapting methods to fit nonacademic organizations, as discussed in Chapter 15.

▪ DESIGNING THE STUDY

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

Study Plan

The process of developing the study plan begins with a one-sentence research question that specifies the main predictor and
outcome variables and the target population. Three versions of the study plan are then produced in sequence, each larger and more
detailed than the preceding one.

Study outline (Table 1.1 and Appendix 1A). This one-page summary of the design serves as a standardized checklist to
remind the investigator to address all the components. The sequence has an orderly logic that helps clarify the investigator’s
thinking on the topic.

Study protocol. This expansion on the study outline usually ranges from 5 to 15 pages. It is used to plan the study and to
apply for IRB approval and grant support. The protocol parts are discussed throughout this book and summarized in Chapter
20.

Operations manual. Studies that enroll participants and collect data also develop a comprehensive manual that includes
specific procedural instructions, questionnaires, data collection forms, and other materials. Even studies that use precollected
data, such as from an electronic health record, must specify exactly how those data will be used. The manual ensures a
uniform and standardized approach to carrying out the study with good quality control (Chapter 18).

The research question and study outline should be written out at an early stage. Putting thoughts in writing leads the way from vague
ideas to specific plans and provides a concrete basis for getting advice from colleagues and consultants. It is a challenge to do
it—ideas are easier to talk about than to write down—but the rewards are a faster start and a better project. Although some early-
stage investigators are understandably impatient to begin their study without taking the time to spell out detailed plans, this often
leads to a poorly conceived project and subsequent delays.

Appendix 1A is an example of a study outline. This kind of outline deals more with the anatomy of research (Table 1.1) than with
its physiology (Figure 1.4), so the investigator must remind herself to think carefully about minimizing the errors that may result
when it is time (Print pagebreak 12) to draw inferences from the study findings. A study’s virtues and problems can be revealed by
explicitly considering how the question the study is likely to answer differs from the research question, given the plans for acquiring
participants and making measurements, and the likely problems of implementation.

With the study outline in hand and the intended inferences in mind, the investigator can proceed with the details of her protocol.
This iterative process includes getting advice from colleagues, drafting specific recruitment and measurement methods, considering
scientific and ethical appropriateness, modifying the study question and outline as needed, pretesting specific recruitment and
measurement methods, making more changes, getting more advice, and so forth.

Trade-offs

Regrettably, errors are an inherent part of all studies. The main issue is whether the errors will be large enough to make the
conclusions of the study unreliable. When designing a study, the investigator is in much the same position as a labor union official
bargaining for a new contract. The union official begins with a wish list—shorter hours, higher wages, better benefits, and so forth.
She must then make concessions, holding on to the things that are most important and relinquishing those that are not essential or
realistic. At the end of the negotiations comes a final, but vital, step: She looks at the best contract she could negotiate and decides if
it is worth signing.

The same sort of concessions must be made by an investigator when she transforms the research question to the study plan and
considers potential problems in implementation. On one side are the issues of internal and external validity; on the other, feasibility.
Compromises are always necessary. But once a study plan has been formulated, the investigator should examine it afresh to decide
whether it adequately addresses the research question and whether it can be implemented with acceptable levels of error. This last
step of deciding whether the study as designed is worth pursuing—the equivalent of the union negotiator determining whether to
agree to the proposed contract—is, unfortunately, sometimes omitted. Or its answer is no, and there is a need to begin the process
anew. But take heart! Good scientists distinguish themselves not so much by their uniformly good research ideas as by their alacrity
in turning over those that won’t work and moving on to better ones.

▪ SUMMARY
1. The anatomy of research is the set of tangible elements that make up the study plan: the research question and its

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

significance, and the design, study participants, and measurement approaches. The challenge is to design elements
that balance feasibility with the need to produce valid study conclusions.

2. The physiology of research is how the study works. The study findings are used to draw conclusions about
measurable phenomena in an accessible population (internal validity) in order to apply those conclusions to other
situations, people, and events (external validity or generalizability). The challenge here is to design and implement a
study plan with adequate control over two major threats to validity—random error (chance) and systematic error
(bias)—and that adds to our current knowledge about the research question.

3. In designing a study, the investigator may find it helpful to consider Figure 1.4, the relationships between the
research question (what she wants to answer), the study plan (what the study is designed to answer), and the actual
study (what the study will actually answer, given the errors of implementation that can be anticipated).

4. A good way to develop the study plan is to begin with a one-sentence version of the research question that specifies
the main variables and target population and expand this into a one-page outline that sets out the study elements in a
standardized sequence. Later, the study plan will be expanded into the protocol and the operations manual.

5. Good judgment by the investigator and advice from colleagues are needed for the many trade-offs involved and for
determining the overall viability of the project.

1 Terms in red are defined in the Glossary.

2 Predictors
are sometimes termed independent variables and outcomes dependent variables, but the meaning of these terms is
less self-evident. Predictor variables can be exposures, risk factors or protective factors, treatments or interventions, or test
results, terms that we use throughout the book when appropriate.

(Print pagebreak 13)

Appendix 1A: Outline of a Study

This is the brief study plan for a project carried out by Michael Jung, MD, MBA, which began while he was an anesthesia resident
at UCSF, mentored by Jina Sinskey, MD. Most beginning investigators find observational studies easier to do, but in this situation a
randomized clinical trial of modest size and scope was feasible. The study results were published in a high-impact journal ( 1 ) .

▪ PERIOPERATIVE VIRTUAL REALITY FOR PEDIATRIC ANESTHESIA

Background and Significance

Perioperative pediatric anxiety is a common and important aspect of the psychological impact of anesthesia and surgery on children.
Prior research has shown that the incidence of high anxiety at induction can be as high as 50% in children who present for surgery
and general anesthesia ( 2 ) . Audiovisual distraction, including virtual reality (VR), has the potential to be a safe, noninvasive,
nonpharmacologic modality to reduce anxiety.

Specific Aim

To conduct a randomized controlled trial to examine the effect of immersive audiovisual distraction with a VR headset on
perioperative anxiety in pediatric patients presenting for elective surgery and general anesthesia.

Methods

Overview of Design

In this prospective, randomized, controlled, parallel group study, children scheduled for elective surgery and general anesthesia will
be randomly allocated to either a VR group or a control (no VR) group. Group VR patients will undergo audiovisual distraction with
VR during induction of general anesthesia in the operating room. A validated instrument, the Modified Yale Preoperative Anxiety
Scale (mYPAS), will be used to assess anxiety.

Study Participants

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

The target population for this research project is children ages 5 to 12 years old presenting for elective surgery and general
anesthesia; the accessible population consists of similar children seen at the UCSF Benioff Children’s Hospital.

Measurements

The predictor variable is the intervention (ie, VR headset or no headset). The primary outcome is perioperative pediatric anxiety
(measured as mYPAS score), which will be ascertained three times: in the preoperative holding area, on entering the operating
room, and during induction of general anesthesia.

Randomization and Blinding

Randomization will use a computer-generated random assignment table, incorporated into a programmed study application (Redcap)
and concealed from the study recruitment personnel. A time-stamped entry including medical record number and consent
completion are entered for randomization. Given the nature of the VR headset (a patient is either wearing it or not), blinding will not
be used.

(Print pagebreak 14)

Data Analysis

The predictor variable is dichotomous: VR headset or no VR headset. The outcome variable (mYPAS score at induction) is a
continuous variable ranging from 0 to 100, which will be compared between groups using a t test. Categorical outcomes will be
analyzed with chi-squared analysis. A significance level of 0.05 will be used.

Sample Size Estimates

Based on a prior investigation of preoperative anxiety in children ages 5 to 12 years old ( 3 ) , mean ± SD mYPAS scores in the
control group will be 30.1 ± 8.4. Recruiting 31 participants per group would yield a power of 80% (β = 0.20) to detect a 20%
difference in mYPAS scores at induction between the two groups.

REFERENCES FOR APPENDIX 1A

1. Jung MJ, Libaw JS, Ma K, Whitlock EL, Feiner JR, Sinskey JL. Pediatric distraction on induction of anesthesia with virtual
reality and perioperative anxiolysis: a randomized controlled trial. Anesth Analg. 2021;132(3):798-806.

2. Kain ZN, Mayes LC, Caldwell-Andrews AA, Karas DE, Mcclain BC. Preoperative anxiety, postoperative pain, and behavioral
recovery in young children undergoing surgery. Pediatrics. 2006;118(2):651-658.

3. Moura LA, Dias IM, Pereira LV. Prevalence and factors associated with preoperative anxiety in children aged 5-12 years. Rev Lat
Am Enfermagem. 2016;24:e2708.

(Print pagebreak 15)

Appendix 1B: Exercises for Chapter 1. Getting Started: The Anatomy and
Physiology of Clinical Research
1. The Early Limited Formula (ELF) study was carried out in two academic medical centers in California with the goal of
encouraging breastfeeding by newborn babies who had lost ≥5% of their body weight in the first 36 hours after birth. In this
randomized clinical trial, the ELF intervention consisted of teaching parents to syringe-feed 10 mL of formula after each
breastfeeding until the onset of mature milk production; control parents were taught infant soothing techniques. The
proportion of mothers who reported exclusive breastfeeding at 3 months to a blinded interviewer was 79% in the ELF group,
compared with 42% in the control group (P = 0.02) ( 1 ) .

For each of the following statements, indicate (a) whether it is an internal validity or external validity inference; (b) whether
you think it is a valid inference; and (c) any reasons why it might not be valid.

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc

 Designing Clinical Research
ISBN: 978-1-9751-7440-8 | 5th_Edition

a. For the women in this study, provision of ELF increased breastfeeding rates at 3 months.

b. Provision of ELF to infants born in a Boston community hospital who have ≥5% weight loss in the first 36 hours will
likely lead to higher breastfeeding rates at age 6 months.

c. Based on the results of this study, an international effort to provide formula to most newborns is likely to enhance
successful breastfeeding and improve the health of the newborns and their mothers.

2. For each of the following summaries drawn from published studies, write a single sentence that specifies the design and the
research question, including the main predictor and outcome variables, and the intended sample.

a. Investigators in Winston-Salem, North Carolina, surveyed a random sample of 2228 local high school students about
their frequency of watching wrestling on television in the previous 2 weeks, and 6 months later asked the same students
about fighting at school and on dates. The adjusted odds of reporting fighting with a date increased by 14% for each
episode of wrestling the students reported having watched 6 months before ( 2 ) .

b. To assess whether the amount of breastfeeding protects women against ovarian cancer, investigators surveyed 493
Chinese women with newly diagnosed ovarian cancer and 472 other hospitalized women, all of whom had breastfed at
least one child. They found a dose-response relationship between total months of breastfeeding and reduced risk of
ovarian cancer. For example, women who had breastfed for at least 31 months had an odds ratio of 0.09 (95% CI: 0.04
to 0.19), compared with women who had breastfed less than 10 months ( 3 ) .

c. To see whether an association between dietary saturated fat intake and reduced sperm concentration in infertile men
extended to the general population, Danish investigators collected semen samples and food frequency questionnaires
from consenting young men at the time of their examination for military service. They found a significant dose-
response relation between self-reported dietary saturated fat intake and reduced sperm concentrations (eg, 41% [95% CI
4%, 64%] lower sperm concentration in the highest quartile of saturated fat intake compared with the lowest) ( 4 ) .

(Print pagebreak 16)

d. Embolic stroke in atrial fibrillation may be preventable by occluding the left atrial appendage. Investigators studied
patients with atrial fibrillation undergoing cardiac surgery for another indication ( 5 ) . They occluded the left atrial
appendage during surgery for half of the patients (randomly selected) and did not for the other half. In the 2379 patients
in the occlusion group, 114 (4.8%) had a stroke or systemic embolism, compared with 168 (7.0%) of 2391 without
occlusion (P < 0.001).

References
1. Flaherman VJ, Aby J, Burgos AE, Lee KA, Cabana MD, Newman TB. Effect of early limited formula on duration and exclusivity
of breastfeeding in at-risk infants: an RCT. Pediatrics. 2013;131(6):1059-1065.

2. DuRant RH, Champion H, Wolfson M. The relationship between watching professional wrestling on television and engaging in
date fighting among high school students. Pediatrics. 2006;118:e265-e272.

3. Su D, Pasalich M, Lee AH, Binns CW. Ovarian cancer risk is reduced by prolonged lactation: a case-control study in southern
China. Am J Clin Nutr. 2013;97:354-359.

4. Jensen TK, Heitmann BL, Jensen MB, et al. High dietary intake of saturated fat is associated with reduced semen quality among
701 young Danish men from the general population. Am J Clin Nutr. 2013;97:411-418.

5. Whitlock RP, Belley-Cote EP, Paparella D, et al. Left atrial appendage occlusion during cardiac surgery to prevent stroke. New
Engl J Med. 2021;384:2081-2091.

mcm999 | 130.113.111.210 | 26 September 2023 13:16:43 Utc