Identifying a research area

As researchers, it’s important to choose an interesting general topic for your research.

• You’re going to be spending a lot of time working on this, so you should choose a topic you would genuinely like to know more about.

In addition, consider if the research area is important.

• Is there likely to be a substantial impact of this research on future research?
• Will it help push forward scientific understanding?
• Does it have translational potential, i.e. the ability to inform positive changes to clinical practice or policy changes that will improve health and quality of life?
• Basically, are you going to be able to convince funders to give you money to work in this area?

Introduction to the example used in this workbook

In this workbook, you will learn how to design and conduct a research project around menstrual disorders using a population-based (epidemiological) approach.

Let’s assume you are broadly interested in what factors are associated with menstrual disorders like heavy periods, painful periods and pre-menstrual symptoms.

Within this broad field, you might be interested in factors that cause or predict variability in these symptoms, or the impact of these symptoms on women’s health and quality of life.

You might also be interested in particular age ranges, such as during puberty or around the menopause.

Activity

1. Search the scientific literature site PubMed using key words related to menstrual disorders.

• A good place to start would be to read a recent review on the topic, if you can find one.

2. Make some brief notes about the research field.

• Think about:
  • Is it easy to find literature on this? Has much research been done before?
  • What aspects of this topic interest you the most?
  • Do you think this is important? Would research have the potential to affect a large population? Would the impact for that population be substantial?

Introduction to literature reviewing

Once you have identified a broad area of interest (in this case, menstrual disorders), it is a very good idea to review the existing literature in much greater detail. This will give you a better understanding of what has already been done, the unanswered questions, and what your research could contribute.

Types of review

Literature reviews fall into two main categories: systematic reviews and narrative reviews. The differences are described in the following table, taken from https://libguides.brown.edu/Reviews/types.

You will find many literature reviews that sit somewhere in between these two categories. Many narrative reviews take a more systematic approach in some areas, for example by describing the search strategy and rationale for including/excluding studies. This helps to reduce bias in the review - the authors can be more confident the review is comprehensive and relevant and they are less able to ‘cherry pick’ studies that confirm their prior hypotheses.

Tip for RED07: for Part 1 of your RED07 project, you will probably want to use a combination of systematic and narrative review approaches. You are not expected to conduct a full systematic review with a rigid pre-defined protocol, but it will help to start with a well-defined research question (see the Designing section) and to take a systematic approach to searching for studies and selecting which ones to include. You should also critically appraise the research and consider the quality when writing your review.

Defining the aim of the review

Often, the range of scientific literature related to a particular topic will be huge. For example, the existing literature on menstrual disorders is likely to include research from discliplines such as sociology, anthropology, psychology, economics, public health, epidemiology, endocrinology, genetics, cellular biology, pharmacology and clinical medicine.

The range of research questions addressed by the existing literature will be similarly broad.

While it can be interesting to see the topic from all these different perspectives, it is impractical to conduct focussed research in all areas simultaneously.

Therefore, when we’re conducting a literature review to guide our research, we should have a clear, focussed idea of the aim of our review.

In the Designing section we will discuss how to develop addressable research questions. That skill is important when planning a literature review as well, but the questions addressed in literature reviews will not necessarily be as narrow as those addressed in a single piece of original research.

In our example, we could take the general area of “menstrual disorders” and narrow it down to “lifestyle risk factors for heavy menstrual bleeding”. Here, we have specified one menstrual disorder (heavy menstrual bleeding), a focus on causes (risk factors) and a further focus on risk factors related to lifestyle (e.g. smoking, alcohol, physical activity) in particular. Therefore, the aim of our literature review would be:

to identify and summarise existing evidence around the lifestyle risk factors for heavy menstrual bleeding

Finding papers

To identify literature for a review in a systematic way, you should first create a search strategy. This means defining what key terms you are going to use to search the literature, and in what combination.

This is where it really helps to have a specific, focussed aim in mind for your literature review.

In our example, we would want to identify studies that mentioned terms related to our exposure/independent variables (smoking, alcohol, caffeine, physical activity, diet) and our outcome/dependent variable (heavy menstrual bleeding).

We would probably only be interested in any studies that mention any of the terms related to heavy menstrual bleeding AND any of the terms related to lifestyle risk factors. In PubMed, that would mean setting up a search like this:

(“heavy menstrual bleeding” OR “heavy menstruation” OR “heavy periods” OR “menorrhagia”) AND (“smoking” OR “tobacco” OR “cigarettes” OR “alcohol” OR “caffeine” OR “physical activity” OR “exercise” OR “cycling” OR “running” OR “jogging” OR “swimming” OR “sport” OR “diet”)

We might also want to search other databases, or search references of included studies to identify any more studies we might have missed.

Defining the inclusion and exclusion criteria

Some of the publications we identify through our literature search will not be relevant to the aims of our review. It helps avoid bias if we define from the outset (before running the search) what we mean by “relevant” by having clear inclusion and exclusion criteria.

In our example, we might want to use the following inclusion and exclusion criteria:

There are other criteria commonly used in literature reviews too, for example, specifying a date range during which the studies must have been published, or excluding papers not written in English or studies that are not peer-reviewed (such as conference abstracts).

Screening titles and abstracts

Screening publications for inclusion in the review is usually a two step process:

1. Screening by title and abstract
2. Screening by full text

We will be able to exclude many studies by simply reading the title and abstract, but some will require more in depth assessment by reading the full text.

For transparency, it is important to make a note of how many studies are excluded, at what stage, and for what reason.

This information can be recorded in a flowchart and included in the Methods section of the literature review.

Activity

1. Search the scientific literature site PubMed using:

• (“heavy menstrual bleeding” OR “heavy menstruation” OR “heavy periods” OR “menorrhagia”) AND (“smoking” OR “tobacco” OR “cigarettes” OR “alcohol” OR “caffeine” OR “physical activity” OR “exercise” OR “cycling” OR “running” OR “jogging” OR “swimming” OR “sport” OR “diet”)

2. Screen the first 10 titles and abstracts using the inclusion and exclusion criteria defined in the previous section.

• How many are excluded? For what reason(s)?
• How many are being included in the next stage of screening (full text screening)?

Reading papers

Reading papers is a skill that you will develop as you get more practice. Watch the following video to learn more about how to read a paper:

Interpretation and critical appraisal

It is vital not to take papers at face value, but to critically appraise their contents.

Critical appraisal involves assessing research to judge its trustworthiness and value.

It allows us to distinguish good quality evidence from opinion, assumptions, misreporting and bias.

Different types of studies have different strengths and weaknesses and require slightly different approaches to critical appraisal. The Critical Appraisal Skills Programme (CASP) have published handy checklists for different study designs: https://casp-uk.net/casp-tools-checklists/

Activity

1. Use the advice from the video to read this paper by Kritz-Silverstein et al. that appears to meet the inclusion criteria for our review on lifestyle risk factors and heavy menstrual bleeding.

• Skim read it first: read the title, abstract and skim the rest of the paper (focusing on the headings). Is it relevant to the review?
• Read it again: read it in more detail this time - what are the research questions? The aims? What did the authors do? Who did they study? What did they find?

2. Use the CASP checklist for cohort studies to critically appraise the paper.

Writing the review

A literature review is not merely a list of papers, their findings, strengths and weaknesses.

Although all those elements are important, a good review is organised into a logical structure, with balanced arguments that allow the reader to fully appreciate the current state of knowledge on a subject. It makes connections between different studies and explains how they relate to different themes or arguments.

A literature review might highlight information such as:

• Trends: whether certain approaches become more or less popular over time
• Themes: questions or concepts that keep coming up
• Debates and contradictions: do some studies provide evidence for a hypothesis and others against it?
• Pivotal publications: are there any influential theories or studies that changed the direction of the field?
• Gaps in knowledge: what is missing from the literature? Are there weaknesses that need to be addressed?

Structure

A good first step in writing a literature review is to outline the structure of the review.

This will be broadly similar to a research paper in that you will likely have sections labelled: Introduction, Methods, Results, and Discussion. However, the Methods section of a narrative review is likely to be small, and the Results and Discussion sections might be merged together.

Methods

In a systematic review, the methods section will be very detailed and could be quite long. In a narrative review, it is likely to be much shorter, including just key information on the search strategy (i.e. what databases were used and what the inclusion and exclusion criteria were).

Within the methods section, or alternatively at the start of the results section, it can help readers if you show a flowchart detailing how many studies were identified in your initial searches and how many remained to be included in the review after screening.

Tip for RED07: It is quite common to see narrative reviews without a methods section, but we do not recommend this for MSc projects.

References

A list of references should follow the review.

When writing any academic work, but especially literature reviews, you are likely to read a large number of papers, some of which you will want to cite. Keeping track of all these references can quickly become difficult.

In addition, different journals have different requirements for how references should be formatted in the reference list and cited in text. If a paper is rejected from one journal, it is common to resubmit it elsewhere. On top of the disappointment of rejection, reformatting every single reference can be a pain!

Therefore, researchers are strongly advised to use reference management software to keep track of their references.

Such software usually includes a plug-in for word processing software such as Microsoft Word, which lets you ‘cite as you type’ - easily allowing you to automatically insert citations into your work as you go. These plug-ins then allow you to automattically insert a list of references at the end of your document in whatever format you choose.

---

Activity

1. Take a look at https://researchguides.dartmouth.edu/citingsources/refs, which includes a comparison of different reference managers.

• If you don’t already have a reference manager installed on your computer, download one!

---

Introduction

Your literature review will have given you an expert and up-to-date understanding of the previous evidence in your research area.

• It might have helped highlight gaps in the literature that your own research could address. It might have highlighted conflicting evidence using different study designs, methodologies or populations.

Whatever your literature review has revealed, you are now in a much better position to plan your own research, beginning with developing a clear, focussed, concise, complex, arguable research question.

Develop a research question

A research question is the question around which you centre your research. It should be:

• clear: it provides enough specifics that your audience can easily understand its purpose without needing additional explanation.
• focused: it is narrow enough that it can be addressed thoroughly in a scientific paper.
• concise: it is expressed in the fewest possible words.
• complex: it is not answerable with a simple “yes” or “no,” but instead requires synthesis and analysis of ideas and data.
• arguable: its potential answers are open to debate rather than accepted facts.

Research questions are essential to the research design process.

Research questions help focus studies by providing a path through the research and writing process. The specificity of a well-developed research question helps writers avoid vague, meandering papers and focus instead on a specific, arguable area.

Steps in developing a research question

The first three steps in developing a research question have already been covered in the Scoping section in relation to choosing a general research area and conducting a literature review of the existing literature:

• Choose an interesting general topic.
  • Researchers should choose a broad topic about which they genuinely would like to know more.
• Is it important?
  • Is there likely to be a substantial impact of this research on future research?
  • Will it push forward understanding in the field? Does it have translational potential, i.e. the ability to inform positive changes to clinical practice or policy changes that will improve health and quality of life?
  • Basically, are you going to be able to convince funders to give you money to do this?
• Do some background research on your general topic.
  • Search a scientific literature database (like PubMed) to see what’s already been done and to help you narrow your focus.
  • What issues are researchers discussing, when it comes to your topic?
  • What questions occur to you as you read these articles?
  • This is where a literature review can be really helpful in helping you shape your research questions.

The next steps are to:

• Start asking questions. Taking into consideration all of the above, start asking yourself open-ended “how” and “why” questions about your general topic.
  • What’s missing from the current literature?
• Evaluate your question. After you’ve put a question or even a couple of questions down on paper, evaluate these questions to determine whether they would be effective research questions or whether they need more revising and refining.
  • Is your research question clear, focused, concise, complex and arguable?
  • Using the skills, data and resources available to you, are you able to address this research question?

Considering research capacity

The extent to which research questions can be addressed is governed in part by the availability of quality, relevant skills, data, facilities and resources.

Collaborating with other researchers and organisations can help increase ‘research capacity’ - your ability to do the research you set out to do.

• You might collaborate with a group with particular skills (such as bioinformaticians), or lab facilities, or access to important datasets.
• Of course, money also helps to buy things like research equipment, consumables, access to facilities and data, as well as pay the salaries of highly-skilled staff.

It is important to bear these practical issues around research capacity in mind when developing research questions.

In the example used in this workbook, you are learning how to design and conduct a research project around lifestyle risk factors for menstrual disorders using a population-based (epidemiological) approach. Therefore, access to an appropriate epidemiological dataset is likely to be an important consideration when planning your research.

Let’s assume you have searched for possible sources of population-based data for your study and you have found the Avon Longitudinal Study of Parents and Children (ALSPAC), a well-known longitudinal prospective cohort study.

The Avon Longitudinal Study of Parents and Children (ALSPAC)

Based at the University of Bristol, ALSPAC, also known as Children of the 90s, is a world-leading birth cohort study.

Between April 1991 and December 1992, ALSPAC recruited more than 14,000 pregnant women into the study and these women (some of whom had two pregnancies or multiple births during the recruitment period), along with the children arising from the pregnancy, and their partners, have been followed up intensively over two decades.

ALSPAC is the most detailed study of its kind in the world and it provides the international research community with a rich resource for the study of the environmental and genetic factors that affect a person’s health and development.

ALSPAC research aims to inform policy and practices that will provide a better life for future generations.

Activity

Now you need to consider whether or not ALSPAC is an appropriate dataset to conduct research on lifestyle risk factors for heavy menstrual bleeding. The availability of data within the cohort will govern what research questions can be addressed using ALSPAC. The ALSPAC variable search tool allows simple searching across variable names and labels to help with identifying what data might be available.

1. Visit the ALSPAC website for researchers and learn more about the cohort

2. Use the search tool to identify variables related to heavy menstrual bleeding

• Try out different terms to ensure you capture all relevant variables.
• Write down the names of all questionnaires in which data on heavy menstrual bleeding has been collected.

3. Use the search tool to find some variables related to smoking, alcohol, caffeine, diet and physical activity.

• Has ALSPAC collected data on all of these lifestyle factors?

Refining research questions

Now that we know what data are available, we can start to think about interesting and important research questions that can be addressed using these data.

Novel research questions are important for driving the field forward, but it is also important to confirm (or refute) findings from previous research using different approaches or by looking at different populations.

It’s also important to consider what we actually want to find out.

• Are we interested in using information on lifestyle to develop better clinical prediction tools for risk of heavy menstrual bleeding?
• Do we think certain lifestyle factors might cause heavy menstrual bleeding, and therefore changes to lifestyle would improve the symptom?
• Or are we just interested in whether lifestyle factors are associated (correlated) with heavy menstrual bleeding at this stage (perhaps because, as is often the case with observational data, we are not able to distinguish cause-and-effect)
• What about the population we want our results to be generalisable to? Is it all women? women of a certain age? Women with certain medical conditions?

It is common for one research project to be comprised of more than one research question. For example, within one project, we might ask questions related to different populations, several exposures (independent variables), several outcomes (dependent variables), and with different motivations (e.g. causation, prediction).

However, it is important to keep the research focused and avoid trying to address too many research questions at once. This can be particularly tricky when working with a dataset such as ALSPAC where so much information is available.

Activity

1. Write down 10 research questions by combining the terms in the table

2. Rank each of your research questions in terms of how feasible they will be to address using ALSPAC data.

Aims

Once researchers have developed their research question(s), the next step is to describe what their project aims to do.

There is usually an overarching aim and then several more specific aims.

In our example, let’s assume we have decided to focus on the following research question:

Research question: Is smoking associated with heavy menstrual bleeding (in women in their early 20s)?

Our overarching aim could be…

Main aim: To assess whether smoking is associated with heavy menstrual bleeding

And our more specific aims…

Specific aims:

1. To explore whether women who are currently smokers are more likely to experience heavy menstrual bleeding compared to women who have never been smokers.
2. To explore whether, in women who smoke, smoking heaviness is associated with higher rates of heavy menstrual bleeding.
3. To assess whether any association between smoking and heavy menstrual bleeding is sensitive to adjustment for key confounding variables (body mass index, socioeconomic status, parity).
4. To assess whether any association between smoking and heavy menstrual bleeding is affected by hormonal contraception.

Activity

1. Assume for a moment that our main aim is actually “to assess whether physical activity is associated with heavy menstrual bleeding”

• What specific aims could your study explore using ALSPAC data?

Hypotheses

Hypotheses are closely related to research aims, but rather than describing what you plan to achieve, they express what you “expect” to find. They often introduce the direction of association/effect that we are expecting.

In our example, our aims, reworded as hypotheses, could be:

Main hypothesis: Smoking is associated with a higher rate of heavy menstrual bleeding

And our more specific hypotheses…

Specific hypotheses:

1. Women who are current smokers are more likely to experience heavy menstrual bleeding compared to women who have never been smokers.
2. In women who smoke, smoking heaviness is associated with higher rates of heavy menstrual bleeding.
3. Adjusting for key confounding variables attentuates the association between smoking and heavy menstrual bleeding.
4. The association between smoking and heavy menstrual bleeding will be attenuated in women on hormonal contraception relative to those not on hormonal contraception.

Objectives

A study’s objectives describe what will actually be done. Again, these are related to the aims, but there is an extra element describing the methodology (how you will achieve the aims).

In grant applications, the objectives are usually presented reasonably concisely as bullet points (as with the research questions and aims), with further elaboration in a separate methods section.

Objectives in our example:

1. To conduct logistic regression analysis on to explore whether mothers in ALSPAC who are current smokers are more likely to self-report that they experience heavy menstrual bleeding than mothers in ALSPAC who have never been smokers.
2. To conduct logistic regression analysis, to assess whether, among ALSPAC mothers who smoke, number of cigarettes per day is associated with higher rates of self-reported heavy menstrual bleeding.
3. To compare logistic regression models to assess whether any association between smoking and heavy menstrual bleeding is sensitive to adjustment for key confounding variables (body mass index, socioeconomic status, parity).
4. To compare logistic regression models to assess whether any association between smoking and heavy menstrual bleeding is affected by hormonal contraception.

Methods

The research methods will obviously be quite different depending on what kind of study you’re conducting (e.g. a lab-based study, a clinical trial, a qualitative study, an observational epidemiological study, etc). However, many of the considerations will be the same.

It is hugely important to really think through these details before conducting the research. This helps prevent bias that can be introduced due to “making it up as you go along”! It is also likely to save a lot of time, money and resources.

Important considerations:

• What study design are you using (e.g. case control, cohort, trial, animal experiment)
  • Do you need ethical approval for this study? (are you likely to get it?)
• Where is your data coming from?
  • If it already exists, how are you getting access to it?
  • If you’re collecting it yourself, how?
  • Who are your participants? How will you recruit them?
  • How will they provide data/samples? What data/samples will they provide?
  • What laboratory techniques will you use to generate data?
  • Is your technique/approach the best and most up-to-date?
  • Is it going to work in this context, using the samples/data you have?
  • What facilities, equipment and consumables do you need?
  • How will you ensure the quality of the data you generate?
• How will you store, prepare and analyse the data?
  • How will you store the data? Will it be safe?
  • How will you prepare the data and derive the variables of interest?
  • What statistical analysis will you conduct?
  • How will you try to minimise statistical bias?
  • How will you try to ensure your findings are generalisable to the wider population?
  • Is the sample size large enough to provide sufficient statistical power?
• Do you have access to the necessary facilities/resources/expertise or do you need to collaborate externally?
  • Who are you collaborating with and how will the collaboration work?
• How much is this going to cost and how much staff time are you going to need?
  • Is this project feasible?

When designing the methods for a study, it really helps to be clear about what the risks and limitations are. Then you can think about ways to mitigate those risks to improve the feasibility of the study.

Activity

1. Think about the methods of our example study on smoking and heavy menstrual bleeding.

• What study design are we using?
• How will we access the data?
• How should we define “heavy menstrual bleeding”?
• Should number of cigarettes per day be treated as a continuous or ordinal variable?
• Should former smokers appear in our analyses, or are we only interested in ever vs never smokers?
• Women have provided questionnaire data at various timepoints, are we using them all? If not, which ones?
• What statistical tests could we perform?

2. Think about the risks and limitations of our study.

• Are there any risks to the participants?
• Are there any ethical issues?
• Are the data any good?
• Are we likely to have sufficient statistical power?
• If we find null results (no association) for aim 1, will the rest of the project be irrelevant?

Writing a grant proposal

Nearly all research requires funding, so a big part of a researcher’s job involves applying for grants from funding bodies.

These funding bodies include charities (e.g. Wellbeing of Women, Cancer Research UK, British Heart Foundation), national government funders (e.g. the Medical Research Council, the Biotechnology and Biological Science Research Council) and international organisations like the European Research Council.

Grant proposals take a long time and a lot of effort to write… and the vast majority are rejected, usually without review. Most researchers will say they’ve had more than their fair share of grant rejections. Even the most highly experienced researchers still face regular rejections. It’s a necessary part of the job and requires a heavy dose of “academic resilience”!

That said, grant writing is a skill that develops with practice and writing a grant, even if it is rejected, is rarely a waste of time. Writing the proposal itself can be hugely beneficial for guiding a researcher’s thinking about a topic.

Examples of grant proposals

Please note that the following are provided just as examples and they will have different sections that you don’t necessarily need for RED07! The first three are “real life” grant proposals that were funded. The final one is my RED07 project from when I was a student on the RED MSc back in 2010… there may be differences in expectations and requirements back then compared to now :)

• MRC New Investigator Research Grant: Prenatal influences on childhood health: what role for mums and dads?
• NIH: Human immune system layering and the neonatal response to vaccines
• NIH: The Semen Enhancer of HIV Infection as a Novel Microbicide Target
• Gemma Sharp’s RED07 project from 2010

Structure

There are some sections of grant proposals that are common to most application forms. It makes sense for researchers to get good at writing these!

• Abstract: a very concise overview of the proposal. It describes the background and rationale, the approach/methods and the likely impact.

• Lay Summary: similar to the abstract, but written for a ‘lay’ (non-specialist) audience. Harder than it might seem!

• Background: sets the scene for the research, describing (briefly) previous evidence in the field and highlighting the gaps that this research project will attempt to fill.

• Research question/hypothesis and aims: these should be focused, clear and concise.

• Experimental design and methods to be used: this should be detailed with headings to break it up so it’s easy for reviewers to read. It often helps to structure this section according to the various aims of the project. Remember to provide a rationale for why certain methods are used.

• Consent and Ethical considerations: a very important section, especially in clinical research or any research that uses humans or animals. For our example study, this section will actually be quite short because all the data already exist. It would be sufficient to say something like “Ethical approval for the ALSPAC study was obtained from the ALSPAC Ethics and Law Committee (IRB00003312) and Local Research Ethics Committees. Informed consent for the use of data collected via questionnaires and clinics was obtained from participants following the recommendations of the ALSPAC Ethics and Law Committee at the time of collection”.

• Justification of resources: this section outlines what the funding will be spent on. Researchers need to ensure this section is comprehensive and nothing has been left out. In addition to staff salaries and costs of equipment, data access and consumables, it is common to also request some funding for staff training and attendance at conferences (travel and subsistence).

• Potential Outcomes and Limitations: it is helpful to explain in this section what the likely outcomes of the research will be. What is the likely impact? How will the findings be used? Will they push the field forward? Will they lead to any changes in clinical practice or policy? Also, what are the limitations of the research (e.g. will self-report data on HMB accurately reflect the clinical definition? does observational data from ALSPAC help us infer causal relationships?)? How are you taking steps to minimise the limitations?

• References: as with most research documents, the grant proposal ends with a list of references, most of which will have been cited in the background section.

Activity

1. Take a look at Cancer Research UK’s guidance on writing for a lay audience

2. Look at the abstract for our example research project below. Try to write the lay summary.

On any given day 800 million women in the world are menstruating and estimates suggest around a quarter (200 million) experience heavy menstrual bleeding (HMB), which is the single most common cause of iron deficiency anaemia in the developed world in premenopausal women. In addition, HMB can lead to fatigue, anxiety, reduced mood and energy levels, with a negative impact on quality of life and productivity. Despite the huge public health burden of HMB, there has been very little research in the area and the risk factors for the disorder are unclear. Consequently, when women present with HMB, the treatment options are pharmacological (e.g. NSAIDs, tranexamic acid, orgal contraceptives, oral progesterone, hormonal IUD) or surgical (e.g. uterine artery embolisation, endometrial ablation, hysterectomy) and very little advice is available on how to prevent HMB through less clinical, lifestyle changes. The study described in this proposal aims to identify whether smoking is a risk factor for HMB by conducting logistical regression analyses using detailed, longitudinal data from ~ 14,000 mothers enrolled in the Avon Longitudinal Study of Parents and Children (ALSPAC). A better understanding of the association between smoking and HMB would inform further research into its use in clinical prediction of HMB risk and the causal role of smoking in the aetiology of HMB. Our findings could be used to help women and healthcare providers better understand the causes of HMB. They could also be encorporated into smoking cessation advice as an additional motivation for women to quit.