How to Write a Study Protocol

8
Lukas B. Moser and Michael T. Hirschmann

8.1 Introduction facilitate the first steps into the unknown territory
of research [4].
This first step of getting the research idea orga- Secondly, in many countries young orthopedic
nized may be difficult for the beginner. However, trainees may not be prepared with a basic scien-
it is also an important step. In the further course tific skillset at medical school. Specific courses
of the research, the young scientist will find out may be rare and often do not focus enough on the
that writing a brief, concise, and comprehensive challenges of junior researchers [3].
study protocol will save time and problems in the Thirdly, from a young researcher’s perspec-
course of the study. If done properly it is defi- tive, it appears to be a waste of time to spend their
nitely one of the most rewarding parts of the valuable time on preparation of a study protocol.
research project as it facilitates later execution of It is the time of publish or perish in which the
the study and the subsequent writing process (see young researcher is driven by the ultimate goal to
Chapter 11.9). get the study done, to present at a conference and
Many junior researchers have difficulties finally get it published.
starting their first research project. This might be Fourthly, there is always a next conference and
due to a number of reasons. submission deadline to meet. This might mislead
Firstly, most of the scientific project ideas young researchers to gather overhasty data in
evolve from problems more senior orthopedic order to write their abstract in time. Please do not
surgeons encounter during their daily practice. feel rushed by upcoming deadlines. This inevita-
The seniors pass the study idea over to the bly decreases the quality of the scientific work,
younger researchers, who often do not fully and the abstract might even get rejected [3].
understand it at first sight as they do not have the Clearly, there is no alternative to writing a
necessary broad and deep orthopedic knowledge. proper study protocol. Therefore, writing a good
Hence, good mentorship from senior orthopedic study protocol is time well spent facilitating your
researchers is of utmost importance and will further study. It is also mandatory for ethical
committee or grant application.
L. B. Moser · M. T. Hirschmann (*)
Department of Orthopaedic Surgery and
Traumatology, Kantonsspital Baselland
(Bruderholz, Liestal, Laufen),
8.2 What Is a Study Protocol?
Bruderholz, Switzerland
A study protocol is the central documentation of
University of Basel, Basel, Switzerland
e-mail: [email protected]; the research project including scientific, ethical,
[email protected] and regulatory considerations. It should provide
© ISAKOS 2019 65
V. Musahl et al. (eds.), Basic Methods Handbook for Clinical Orthopaedic Research,
https://doi.org/10.1007/978-3-662-58254-1_8
66 L. B. Moser and M. T. Hirschmann

detailed information on planning and execution Every research protocol follows a similar
of the research project. The study protocol serves basic structure. However, it is important to know
as a comprehensive guide and also represents the that the structure of a research protocol might
main document for external evaluation of the vary from one to another study. This is due to the
study (e.g., ethical committee, grant authorities). fact that study protocols written for ethical com-
However, the purpose of the study protocol is mittee approval might be slightly different than
to give a concise description of the study idea, protocols written for grant authorities. It is also
plan, and further analysis. The writing style mandatory to follow exactly the instructions
should be brief and concise. It needs to be easy to given by the grant authority or local ethical com-
understand even for laypersons without a medical mittee [9].
background. Generally, the study protocol can be divided
Think of a study protocol as a recipe, which into the following parts:
should enable the cook/reader to prepare an iden-
tical meal [9]. • Title page
• Background
• Study objectives (aims)
• Hypothesis
As a study is often the result of an interdis- • Material and methods
ciplinary collaboration and involves a vari- –– Study design
ety of different professions such as –– Subjects
clinicians, scientists, or statisticians, all –– Sample size
groups should be involved at this early –– Study procedures
stage. Significant input from all groups –– Outcome instruments
involved is necessary. –– Data collection
• Data management
• Statistical analysis
• Ethical considerations
A well-written study protocol is necessary for: • Time points and timeline
• Conflict of interest
• Application for ethical approval at the local
• Insurance
ethical committee.
• Reference section
• Application for a scientific grant at grant
• Annexes
authority.
• It helps to structure and define the study idea To help getting started, the researcher should
in the discussion with the collaborators. consider the following questions to be answered:
• It prompts rethinking the study plan and
reveals possible problems and obstacles at an • What is the clinical problem to be addressed
early stage. by this study?
• Clearly defines the responsibility of each • What is already known about the problem and
researcher involved. topic?
• Defines a budget and funding plan for the • What is the study design in the study?
study. • What are the inclusion and exclusion criteria?
• Defines a clear timeline for each step of the • What study subjects are investigated?
study. • What are the outcome instruments chosen?
• Helps to monitor the milestones and study • What are the primary or secondary endpoints
progress. of the study?
• Facilitates writing of the later scientific • What are the interventions?
article [14]. • What is the experimental setup?
8 How to Write a Study Protocol 67

• How is data collection organized? ence is really needed to lead the reader to the
• How is the data processed and analyzed? purpose of the study. It also helps to stay
• What are the statistical methods? focused [14].
• Are there any ethical problems to be It is of utmost importance to clarify why the
considered? researcher is conducting the study. The researcher
• What is the timeline? should make the reader understand why the
• What milestones are set? research project is planned and what the original
• How is funding organized [22]? study idea is. As this requires a good understand-
ing of the current knowledge in this field, a sys-
Having specific answers to the aforemen- tematic review of the current literature should be
tioned questions helps to fill every part of the performed. It facilitates the summary of current
study protocol. Hence, we can move on and get it evidence and allows to put the study idea into the
started step by step. broader scientific context [16]. The red thread of
the background section is to start from the eagle’s
perspective and come closer to the topic by every
8.3 Title Page sentence written [9].
Defining the research question is definitely the
The title page contains the most important infor- heart of the protocol. The research question pro-
mation at a single glance. Here, the researcher posed should be able to fill an existing gap of
should give the full title of the study along with the knowledge. Ideally, it should represent the logi-
names and affiliations of all people involved [4]. cal consequence of the background explained.
The project title needs to be wisely chosen as The question asked here should be as precise as
it should closely reflect the content of the study. possible [16].
Keep it brief and concise. The title should not This part decides the fate of the study. Clearly,
pose a question, but summarize the main objec- if the researcher is not able to give a proper study
tive including the study type such as “random- question and explain the hypothesis here, better
ized controlled trial.” It might also mention the reconsider and do not waste the personal resources
subjects to be investigated [14]. and time for this study [14] (Tables 8.1 and 8.2).
The affiliations of each study team member A major pitfall of study protocols is that many
need to be complete. For every member of the do not have a single study question and others
study team, the contact information including just have too many. There might be several study
e-mail address and telephone and fax number
should be given here. Table 8.1 The mnemonic FINER will help the researcher
If applicable state the study sponsor, in the with identification of a good research question and study
case the study is supported by a company or grant idea [7]
authority [4]. F Feasible in terms of recruitment, expertise, funding,
resources, and scope
I Interesting to the scientific community
N Novel, providing new findings and confirming or
8.4 Background rejecting previous findings
E Ethical approval should be possible
The background of the topic should be clearly R Relevant to scientists, clinicians, and future research
described. Meticulously guide the reader into
the topic while avoiding irrelevant informa- Table 8.2 The PICO acronym helps to formulate a
tion. Do not spend more than two pages on the ­testable question [15]
background information. As a rule of thumb, P Patient, problem
also limit the scientific articles for reference I Intervention or exposure
to less than 20–30. Only describe the most C Comparison
important ones here. Always ask if this refer- O Outcome
68 L. B. Moser and M. T. Hirschmann

questions of interest with regard to the research known to be possible causes of patellar overstuff-
project that need to be answered. However, the ing inducing a higher patellar BTU, these possi-
researcher should stay focused on the most ble bias parameters will be calculated and
important research question. Secondary research compared between the two groups.
questions are often of explorative nature, because The secondary purpose is to compare postop-
sample size calculation is based on the primary erative functional results measured at 1 and
research question. Do not overload the project 2 years postoperative.
with too many research questions. The researcher An exact definition of outcomes is mandatory to
will then lose the red threat and the reader will be standardize the outcome measures. Note the time
confused. The researcher also has to consider that of assessment and the unit of measures. If previous
each research question requires a hypothesis. definitions or validations are preferred, do not for-
Defining the research questions and objectives in get to quote the references. If the researcher com-
advance helps the researcher to avoid reporting pares outcomes, (s)he needs to state an overall goal
significant results more likely than negative of the comparison in the protocol. The reader
results (outcome reporting bias) [16]. should understand if the goal was to show superior-
ity, equivalence, or non-­inferiority [16].
Sometimes, outcomes are not easy to deter-
8.5 Study Objectives (Aims) mine. Some outcomes might be subjective (e.g.,
pain); others are more objectively measurable
After performing the literature review with a (e.g., range of motion).
focus on the research question, the researcher Surrogate endpoints are alternative endpoints
needs to define the objectives, outcomes, and that are faster to assess than long-term clinical
hypothesis. A clear definition of these factors is endpoints [8]. The effect of the intervention on
recommended [9] (Table 8.3). the surrogate endpoint has to correspond to the
Choose the study objectives wisely and effect on clinical outcome. However, this effect is
restrictively. If the researcher decides to use more often difficult to predict, and therefore surrogate
objectives, (s)he should distinguish between pri- endpoints should only carefully be used [12].
mary and secondary objectives or outcomes. Furthermore, the researcher needs to identify
Altogether, do not choose more than three to five all potential confounders. Confounders are
objectives [4]. defined as additional factors that distort the effect
Practical Case Example of the treatment or exposure on a predicted out-
A clinical study investigating if the femoral come. However, they have an impact on both, the
prosthetic design in total knee arthroplasty exposure and outcome, while not being part of
(TKA) influences the patellar loading and func- the causal pathway. Confounding factors might
tional scores in TKA with unresurfaced patellae lead to the situation that falsely a correlation
The primary purpose of the study is to investi- between treatment and outcome might be seen or
gate if the femoral shape of the different TKA a true relationship is hidden. Therefore, con-
models, (group P) and (group A), influences the founders should be carefully considered. Ideally,
BTU distribution pattern at SPECT-CT in TKA confounders are equally distributed in random-
with unresurfaced patellae. Since femoral TKA ized controlled studies. Be aware that this is not
malrotation, patellar height, and TT-TG are the case in observational data [18].
Table 8.3 It helps when the researcher defines the objec-
tives of the study with regard to the SMART criteria [6]
Clinical Vignette
S Specific
M Measurable Age being a confounder in a study investi-
A Achievable gating the association of physical activity
R Realistic and knee pain
T Time-related
8 How to Write a Study Protocol 69

identified the subjects or patients, what were the

Younger individuals tend to be more inclusion and exclusion criteria (study popula-
active and have a lower risk of knee pain. If tion), and what is investigated (study procedure).
the proportion of young people being more Explain each step of the study procedure meticu-
active with less pain is not equally distrib- lously. Only then an independent reader is able to
uted, the association between activity and follow and reproduce the study [4].
knee pain might be overestimated [21].
Vavken et al. investigated the consider-
The methods should include the following
ation of confounders in 126 controlled
aspects: design and setting, subjects, sam-
studies published in journals with a high-
ple size, description of the study procedure,
impact factor. Although 16% of the studies
data collection, data management, and
discussed confounding factors, they did not
finally the statistical tests to be used.
adjust their subsequent analysis. Only 1/3
of the authors controlled for confounding
factors [21]. However, confounders need to
be identified in the study protocol. Only 8.8 Study Design
then the later analysis can be adjusted [2].
Precisely define the study type and design. One
can differentiate between research on primary
8.6 Hypothesis and secondary data. Whereas research on pri-
mary data means the researcher is performing the
The hypothesis is based on the supposed relation actual study (e.g., clinical, experimental, and epi-
between variables. State the hypothesis in the demiological studies), research on secondary
null form. The null hypothesis says that there is data describes the process of analyzing studies
no relationship between variables and the that have already been performed (e.g., system-
researcher is going to challenge that statement. It atic review, meta-analysis).
represents the contradictory of what to expect. Chapter 11.9 provides a comprehensive list of
Statistical testing will reveal the probability that all possible study types. Although in practice
the null hypothesis is true or false. The alternative often difficult, the researcher should always aim
hypothesis is recognized when the null hypothe- for the study type providing the best scientific
sis is rejected. It represents the outcome the evidence and quality level.
researcher would expect [14].

8.9 Subjects

8.7 Material and Methods The study subjects or patients included need to be
characterized in detail. Including a flow chart
The methods section needs to explain how the showing the recruitment, screening, inclusion,
hypothesis is tested. It is of utmost importance to and exclusion criteria will help the reader to
find the optimal methodological approach for the understand this process (Fig. 8.1).
study. After a first draft, the researcher often has to The eligibility criteria need to be mentioned
critically revise the methodology considering along with all inclusion and exclusion criteria.
input from all collaborators [9]. An exact and pre- Inclusion criteria determine which subjects or
cise description of the methods used is the most patients are going to be included into the study.
important part of the study protocol. At the begin- All other criteria limiting eligibility are consid-
ning state the study design and then continue with ered as exclusion criteria [19].
the description of the study subjects. The reader Explain how and why the researcher decided
must be able to understand how the researcher to choose the sample size used in the study. This
70 L. B. Moser and M. T. Hirschmann

Does subject 8.11 Study Procedures

meet eligibility?

Yes
In this section, the researcher should clearly describe
what is going to be done with the s­ ubjects, how, and
Invite subject to
participate in study when. Provide a detailed description of the study
procedure with the planned interventions.
Consent process
successful?
8.12 Data Collection
Yes
Subjects enrolled?
Describe clearly who and how to obtain the data,
Fig. 8.1 Flowchart showing enrollment process
how to record it, and how long it will be stored
[4]. Name the locations where the data is going to
be collected, relevant events regarding recruit-
needs to be based on a sample size calculation.
ments, exposer, or follow-up [20].
Keep in mind that an inclusion of vulnerable pop-
State the data-collecting instruments (e.g.,
ulation (children, cognitively impaired adults)
electronic forms or paper forms). Explain the
requires special justification [5].
validation procedure for the instruments used.
Provide a detailed and unambiguous list of the
For outcome instruments, provide original refer-
eligibility criteria as in the example below:
ences. If the researcher is going to develop a
Enclose a statement which states that the
novel type of data-collecting instruments, as a
participation of the patient is entirely volun-
checklist or a questionnaire, the researcher needs
tary, and an unexplained withdrawal is possible
to provide all information. Do not forget to annex
at any time having no impact on patient man-
the original document to the study protocol.
agement [16].
Practical Case Example
Practical Case Example
This is a retrospective case series study. No
Any subject is entitled to withdraw from this
additional examinations other than the clinical
clinical investigation for any reason without obli-
routine will be conducted. All data is stored in a
gation or prejudice to further treatment.
centralized electronic database at the orthopedic
Declare how the patient consented to partici-
research unit (…). The data has been already col-
pate in the study.
lected in our ethical committee approval registry.
Practical Case Example
The patient shall give his written consent to
participate in the clinical documentation. The 8.13 Data Management
signed consent form remains within the clinical
study case file. The researcher needs to make a comprehensive
statement on the data management including data
entry and monitoring. Describe how the
8.10 Sample Size researcher plans to ensure privacy protection
(e.g., anonymization) and how long the data is
Choosing the appropriate study sample is para- stored. Declare who is able to access the data,
mount for successful research. The study sam- and state that all involved people underline medi-
ple ensures the sufficient power of the study and cal confidentiality. Add the statement by the IEC
might allow conclusive inferences [16]. and regulatory authorities for permitting data
Consider that having a small sample size might access, if applicable [16].
not be powerful enough to prove a difference, Mention how to introduce the data into the
although true difference actually exists. database. Describe what programs the researcher
However, a too large sample size might be a will use and how to perform data analysis. The
waste of resources [14]. best approach is to make a data analysis plan
8 How to Write a Study Protocol 71

where the researcher describes which measure- 8.15 Ethical Considerations

ments were done with which variable and which
statistical tests to apply. Also, record how to deal This part is particularly important for the ethical
with missing data [9]. committee. It will be carefully read; in particular the
ethical risks will be discussed between the members
of the ethical board. Present a risk-­benefit assess-
8.14 Statistical Analysis ment providing information on potential benefits,
risks, or inconveniences for the individual study
A researcher should have basic knowledge of sta- subjects. If necessary, justify the inclusion of vul-
tistics [11]. However, consulting a professional nerable populations. Include a statement that the
statistician might help to improve the quality of researcher conducts the study according to the study
the research methodology. At best the statistician protocol and good clinical practice (GCP) [16].
is already involved in the process of study plan-
ning, as the researcher needs to translate the
research question into a statistical problem. All Keep in mind that any changes of the study
statistical relevant information must be included protocol need to be amended to the ethical
in the sufficient level of details: exploratory or committee and the regulatory authorities.
descriptive statistics, level of significance, type
of outcome, effect measures including confi-
dence intervals, type of sample (unpaired or
paired), data distribution (not normally or nor- Practical Case Example
mally distributed data), and the statistical soft- The study will be performed in accordance
ware used [16]. with the declaration of Helsinki and the direc-
Just naming the tests that were used is not tives for good clinical practice (GCP) standards.
enough. The researcher needs to state why a test Ethical approval will be obtained from local ethi-
was chosen and which test for which purpose [4]. cal committee in accordance with national law.
Practical Case Example
A clinical study investigating if the femoral 8.16 Time Points and Timeline
prosthetic design in total knee arthroplasty
(TKA) influences the patellar loading and func- Creating a table including the time points and
tional scores in TKA with unresurfaced patellae timeline is useful for several reasons. It forces
Descriptive statistics (means, medians, the researcher to set a time frame and think
quartiles, ranges, standard deviations) will be about important deadlines and targets. The
performed to assess the demographics of the reader understands the sequence of events at a
patient population. Alignment and TKA com- single glance, and the timetable shows the
ponent position in all planes (sagittal, coro- reviewer that the project is expected to be com-
nal, and rotational) are noted in degree. Mean pleted in the foreseeable future [13] (Tables 8.4
and absolute relative BTU of corresponding and 8.5).
quadrants of the two groups will be compared
with a t-test.
Nonparametric Spearman’s correlation coeffi- Table 8.4 Practical case example of time points
cients will be used to correlate the patella height, Writing the study protocol January
the lateral tilt, and the components’ alignment Data collection March–
August
measurements with the intensity of the tracer
Data analysis September
uptake in each area of interest. Postoperative Discussion of results among research October
KSS scores of the two groups at 1 and 2 years staff
postoperative will be compared with a t-test. Writing of scientific article November
All data will be analyzed by an independent Submission to peer-reviewed journal December
professional statistician.
72 L. B. Moser and M. T. Hirschmann

Table 8.5 Practical case example of a timeline

Jan Feb Mar Apr May Jun Jul Aug Sep Oc t Nov Dec
Writing the study
protocol
Data collection
Data analysis

Discussion of results
among research staff

Writing of scientific
article

Submission to peer-
reviewed journal

8.17 Conflict of Interest sate serious adverse events (as injuries) the
patients suffer during their participation.
Do not forget to disclose the conflicts of interests, However, rules and requirements of trial insur-
such as nonfinancial or financial relationships ance vary from country to country. Therefore, the
with the industry. If applicable, state the study researcher needs to check the local insurance
sponsor. Declare how the sponsor contributed policy [10].
and what benefits the researcher will obtain by Practical Case Example 1
this agreement [4]. The sponsor will secure and maintain in full
These disclosures facilitate the identification force and effect, throughout the duration of the
of possible sponsor bias. Most journals require a clinical investigation, a clinical trial insurance
disclosure form of all contributing authors. A which was required in line with national regula-
short statement of relevant information will be tions and which will be evidenced by an insur-
listed on the paper in abbreviated form [17]. ance certificate.
Practical Case Example 1 Practical Case Example 2
One of the authors (XY) is a consultant for No subject insurance is required as it is a ret-
(company). rospective study and the patient is not put at any
Practical Case Example 2 risk or has to show up for follow-up.
The study was supported by a financial grant
from (company). The external funding source did
not have any influence on the investigation. 8.19 Reference Section
Practical Case Example 3
The authors of this paper declare no conflict of The reference section should be organized as
interest. in a scientific paper (see Chapter 11.9).
Provide original references, and number them
consecutively or in the order requested. For
8.18 Insurance every statement made, a citation should be
given. It is important that only articles are
The sponsor of the study is obliged to provide cited which really add to the study protocol.
clinical trial insurance. If the researcher is con- There should not be more than 30–40 refer-
ducting a clinical study, (s)he needs to compen- ences in the reference list [4].
8 How to Write a Study Protocol 73

8.20 Annexes 7. Doran PC. Understanding dentures: instructions to

patients. J Colo Dent Assoc. 1981;59:4–8.
8. Fleming TR, DeMets DL. Surrogate end points in
In particular for major grant applications, annexes clinical trials: are we being misled? Ann Intern Med.
could be added at the end. These might include 1996;125:605–13.
informed consent sheet, study questionnaire or CRF, 9. Fronteira I. How to design a (good) epidemiological
observational study: epidemiological research proto-
approval of ethical committee if already obtained, col at a glance. Acta Medica Port. 2013;26:731–6.
and CV of the principal and co-­investigator [1]. 10. Ghooi RB, Divekar D. Insurance in clinical research.
Perspect Clin Res. 2014;5:145–50.
Take-Home Message 11. Gilmore SJ. Evaluating statistics in clinical trials:
making the unintelligible intelligible. Australas J
• Writing a brief and comprehensive study pro- Dermatol. 2008;49:177–84; quiz 185–176.
tocol is always the first step of your research 12. Huang Y, Gilbert PB. Comparing biomark-
project. ers as principal surrogate endpoints. Biometrics.
• It facilitates later execution of your study and 2011;67:1442–51.
13. Kanji S. Turning your research idea into a proposal
the subsequent writing process. worth funding. Can J Hosp Pharm. 2015;68:458–64.
• Make sure that your study protocol includes 14. O’Brien K, Wright J. How to write a protocol. J
all relevant information on your study idea, Orthod. 2002;29:58–61.
plan, and further analysis. 15. Richardson WS, Wilson MC, Nishikawa J, Hayward
RS. The well-built clinical question: a key to evidence-­
• Furthermore, as a comprehensive guide, it based decisions. ACP J Club. 1995;123:A12–3.
represents the main document for external 16. Rosenthal R, Schafer J, Briel M, Bucher HC, Oertli
evaluation of your study. D, Dell-Kuster S. How to write a surgical clinical
• The idea of the study protocol is to convey to research protocol: literature review and practical
guide. Am J Surg. 2014;207:299–312.
any individual the purpose of the study and the 17. Sade RM, Akins CW, Weisel RD. Managing
probable impact of the project. conflicts of interest. J Thorac Cardiovasc Surg.
2015;149:971–2.
18. Skelly AC, Dettori JR, Brodt ED. Assessing bias: the
importance of considering confounding. Evid Based
References Spine Care J. 2012;3:9–12.
19. Van Spall HG, Toren A, Kiss A, Fowler RA. Eligibility
1. Al-Jundi A, Sakka S. Protocol writing in clinical criteria of randomized controlled trials published in
research. J Clin Diagn Res. 2016;10:ZE10–3. high-impact general medical journals: a systematic
2. Babyak MA. What you see may not be what you sampling review. JAMA. 2007;297:1233–40.
get: a brief, nontechnical introduction to overfit- 20. Vandenbroucke JP, von Elm E, Altman DG, Gotzsche
ting in regression-type models. Psychosom Med. PC, Mulrow CD, Pocock SJ, et al. Strengthening the
2004;66:411–21. Reporting of Observational Studies in Epidemiology
3. Bando K, Sato T. Did you write a protocol before (STROBE): explanation and elaboration. PLoS Med.
starting your project? Gen Thorac Cardiovasc Surg. 2007;4:e297.
2015;63:71–7. 21. Vavken P, Culen G, Dorotka R. Management
4. Barletta JF. Conducting a successful residency of confounding in controlled orthopaedic trials:
research project. Am J Pharm Educ. 2008;72:92. a cross-­ sectional study. Clin Orthop Relat Res.
5. Brody BA, McCullough LB, Sharp RR. Consensus 2008;466:985–9.
and controversy in clinical research ethics. JAMA. 22. Warren MD. Aide-memoire for preparing a protocol.
2005;294:1411–4. Br Med J. 1978;1:1195–6.
6. Doran G. There’s a S.M.A.R.T. way to write man-
agement’s goals and objectives. Manag Rev.
1981;70:35–6.