Research methods for Public Health
Aperçu des sections
-
Competences (learning outcomes) to be gained from the Research Methods course.
Demonstrate the ability to:
- Identify an important public health issue facing your community
- Perform a systematic review of the literature to provide the background to your research
- Identify the methods for performing a research project relevant to the problem you have identified
- Prepare a protocol to perform the research project that you have identified
-
Note: This course on research methods has been adapted from the Dissertation module previously used in the Peoples-uni master's course in public health. Thanks to those who contributed resources, in particular Professor Petra Buttner. The course is designed for people to work through it themselves, although it would be suitable to include in a formal course, or to form the basis of discussions with a mentor or supervisor.
How to navigate the course
Each section comprises a set of resources that we think you will find interesting - click on the collections of resources in each section. There is a forum in each topic for reflection.
We encourage you to reflect on what you have learned or comment on the course. When you click on the hyperlink in each topic labelled reflection, you will be able to add a new topic or respond to a previous one. You may want to share your learning from this and other readings, comment on the topics from your own experience, comment on others' posts, or provide feedback on how we can improve the content and presentation.
In the final section you will see that you can gain a Certificate of Completion - the requirements for this are to post a reflection in each section.
This work is licensed under a Creative Commons Attribution 4.0 International License. -
IDENTIFY YOUR PUBLIC HEALTH PROBLEM
Step 1:
1. Identify the general health problem you would like to study (it should be relevant to your population)
Step 2:
Read around the subject enough to help you to identify which aspect you want to study further (prevalence/causes/healthcare etc). This will help you develop the details of the project you plan, and the literature review that you will need to do to inform the project. We suggest that you take the following approach:
(1) chose an issue that fascinates you; and which you have identified to be a real health problem facing your community;
(2) for a start, read around the problem holistically and learn about the various aspects of the problem, do not rush into one aspect straight away but come to it through a a broader study and understanding of the problem. You will need to do an initial/broad review of the literature and any other sources to give you an overview of the problem.
This part should conclude with the identification of the Public Health aspect you want to study further.
-
Consulter Écrire des messages de forum : 1
Reflect on:
What is the general Public Health problem you have decided to study?
- Have you considered the nature of the problem and if the context is relevant to you?
Which particular aspect of the study have you decided upon?
- Have you considered all aspects of the problem - extent, causes, possible solutions and the reasons for prioritising this aspect?
- Have you read some references and considered the results?
- Can you identify a focused research question for the literature review that would follow?
-
-
SYSTEMATIC REVIEW OF THE LITERATURE
Once you have broadly agreed a particular aspect of the Public Health Problem you wish to study, you need to undertake a detailed literature review to find out what is already known about that aspect. To do so, you should develop a search strategy, conduct a literature search (using appropriate MeSH terms) and screen / sort citations, read / critically review the literature.
In undertaking your literature review, you may find the PICOT scheme helpful in developing your ideas (but be aware that PICOT may not apply for all studies and there are other schemes you might prefer to follow, the important thing is to do this exercise systematically and use a well-established/proven framework or methodology – and ideally you should reference this):
Asking the Public Health question—PICOT
Population. To which population or part of your population does the question apply?
Intervention or risk factor (exposure) or prevalence. For an intervention: is the intervention aimed at the individual or the population (e.g. the environment)? At what stage of the disease process are you aiming (e.g. primary or secondary prevention)? At what stage of development is the intervention? Is there a policy option available? What type of intervention is it (e.g. health care, health care organization, environmental modification, community-wide education etc.)?
For a risk factor: can the risk for the population—rather than for the individual—be measured? What sort of risk factor is it (e.g. physiological variable, individual behaviour, social attribute, environmental exposure etc.)?
For a prevalence question: what measurements are feasible?
Comparison. What comparison group are you planning (e.g. before/after study, control group), if relevant?
Outcome. Which outcome measure are you going to have (risk factor level, death, hospitalization, quality of life, cost-effectiveness etc.)? Over what time period is the outcome to be assessed, and how does this relate to the policy cycle length? Will estimates of cost be made?
Type of evidence required. Is this an estimate of burden of illness aiming to quantify prevalence, risk or benefit? Is this to examine the impact of risks or benefits on health inequalities (such as an ‘equity audit’), or the impact of a health policy initiative? Is this an economic evaluation?
RESEARCH SKILLS PROGRAM PART 1
The first part of the research skills program (RSP) will introduce you to some basic ideas about research in general and about systematic literature reviews.
Read the following documents and then take your time to work on the exercise - you will find that time spent on the exercise will be most helpful.
-
Consulter
Please read this document which sets the scene for thinking about, and getting started on, research. It is one of the key background documents for this part of the Research Skills Programme
-
Please read this document which outlines the steps for performing a literature review. It is one of the key background documents for this part of the Research Skills Programme, and will be very helpful in preparing you to work on your literature review
-
Working through this exercise will be very useful to consolidate what you have seen in the resources above, and prepare you to start on your literature review
-
Exercise article to be assessed using CASP.
-
Écrire des messages de forum : 1
This is for you to reflect on any issues relating to the information in the Research Skills Programme 1.
Have you read each of the attachments?
Does this help you think about the study you have in mind?
-
Here is an example of a systematic review. Please focus on the Methods section. This is rather complex, but it a gold standard systematic literature review.
-
Determinants of antenatal care utilisation in sub-Saharan Africa: a systematic review is an excellent example of a systematic literature review. Probably more extensive than you need, but please look through it for inspiration.
-
Following the literature review, you will want to synthesie what you have found. This discussion will help decide if a meta-analysis is warranted beyond making your synthesis.
-
Écrire des messages de forum : 1
This is to encourage you to reflect on how you would perform a literature review. You might want to consider if you can perform the following tasks on a review:
Develop a protocol for the review which has the following features:
- a clear research question and framework, such as PICOT, with all terms defined and key aspects of PRISMA and CASP guidelines incorporated.
Perform the literature review with the following features:
- Several sources explored (eg medical literature databases, appropriate non-medical databases, grey literature, specific journals).
- Varied search strategy for each database as needed. Use of MeSH on non-MeSH headings or combination as needed.
- Use of filters as required.
- Save and list searches and search strategies by source.
Sort citations:
- Exclude duplicates.
- Separate full-text and non-full text.
- Exclude non-relevant papers (and give reasons).
- Prioritise remaining papers (and give reasons).
- Group remaining papers into meaningful sub-topics.
- Create a flow-chart which summarises results (numbers/types of study).
- Organise citations (using traditional citation rules) and create list of references.
Read and critically review each paper.
- Identify evidence type (where relevant) and create table summarising study designs of included articles.
- Assess the quality of the studies and identify biases or other factors.
- Extract relevant statistics from relevant papers.
- Create summary tables (by sub-topics, include all articles in the review, cite first author, year, country, type of study, sample size, main results and critical comments, and relevance to your proposed study population.
Summarise and synthesise.
- Summarise critically.
- Synthesise the evidence (by study type, source and conclusions if relevant).
- Draw conclusions of the relevance to your research topic.
Identify gaps in knowledge.
- What gaps remain following the review.
- How can these gaps lead to the project you plan to conduct.
Identify clear research question for your project, informed by the literature review.
- Is there a plausible hypothesis.
- Can you quantify an expected result.
- Can you identify a study population and a proposed intervention or study factor.
-
-
Sources of full text articles freely available online:
If you have access to a local library, you may be able to access online journals through there. If not, here are several ways to access full text articles for free (note this list is likely to increase as time goes on):
Global Index Medicus provides worldwide access to biomedical and public health literature produced by and within low-middle income countries. The main objective is to increase the visibility and usability of this important set of resources. The material is collated and aggregated by WHO Regional Office Libraries on a central search platform allowing retrieval of bibliographical and full text information. Searchable and should be very valuable.
HINARI http://www.who.int/hinari/en/ has a search facility and you can select various free sources at http://extranet.who.int/hinari/en/journals.php. Note from HINARI: 'If you are not associated to any institution in your home countries, we strongly suggest you approach to registered institutions receiving walk-in users. You can check the list of Registered Universities and Professional Schools at:http://www.who.int/hinari/faq/registration_and_eligibility_registered_universities/en/index.html.
Please explore the Directory of Open Access Journals (http://www.doaj.org). Also the platform for Open Access distribution of articles in journals published in low-income countries, Bioline International (http://www.bioline.org.br).
Highwire has a large number of free articles at http://highwire.stanford.edu/lists/freeart.dtl. Free Medical Journals also has a large list http://www.freemedicaljournals.com/. A simple Google search (free full text articles online) will identify many other possible sources. Pubmed also has links to a number of free full text articles - see this link for more details http://www.nlm.nih.gov/docline/freehealthlit.html.
The Cochrane Library has also now moved to make much of its material open access - see https://www.cochranelibrary.com/ (access may depend on which country you are in and trying to access the material).
There is also has a directory of open access journals at http://www.inasp.info/en/training-resources/open-access-resources/
MedInd http://medind.nic.in/ has a search facilities (mainly Indian journals).
Some more useful tips http://www.tech-recipes.com/rx/6173/how-to-access-subscription-journals-and-full-text-articles-for-free/.
Unpaywall is a free web-browser extension that hunts for papers in more than 5,300 repositories worldwide, including preprint servers and institutional databases. Check out: http://www.nature.com/news/unp
aywall-finds-free-versions-of- paywalled-papers-1.21765 Dimensions also allows you to find possible open source versions of papers - you can search through here https://app.dimensions.ai/discover/publication?and_facet_open_access=True
-
Taking into account the results of the literature review, and the gaps that you can fill in your own research project, you are ready to start to plan your study. Before you start, here are some resources to help with the possible methods you can consider.
Here are the major headings you need to consider in designing a research project.
1. Define the Research Question.
A research question is a question that a study is designed to answer. It guides all other subsequent tasks in the research process and at the end of the project, after data collection and analysis you should be able to answer your question.
The purpose of the research question includes: Organising the study and giving it direction and coherence; Delimiting the study, showing its boundaries; Providing a framework when you write up your research; Keeping the researcher focused; Pointing to the hypothesis, methods and data that will be needed.
Characteristics of a good question include being: Ethical; Novel; Interesting; Relevant; Feasible.
NOTE: the above information on the research question is taken from the Global Health Training Centre and you may want to look at The Research Question through eLearning courses (you will have to register, but it is free to access)
2. Develop the hypothesis. This will be based on the gaps in knowledge you identified in your literature review. In the Research Skills Programme part 2 below, you will find the folder Research hypothesis revisited which contains which provides the way to develop a research hypothesis and which gives two simple examples from previous students.
NOTE: For quantitative studies you will need a falsifiable research hypothesis; i.e. a hypothesis in which you specify the expected outcome in numerical terms. For qualitative studies a proper research question/topic is sufficient. Qualitative studies do not work with hypotheses.
3. Pick your study design. Please look at these: Overview of research designs will help if you are planning a qualitative study.
NOTE: For every research question/hypothesis there is an ideal study design. Please check out RSP_05 The right research design for your research question.
You might want to review the WHO book "Basic Epidemiology" by Bonita, Beaglehole and Kjellstroem (download pdf here).
4. Select your sample to study.
NOTE: The target population is the population to whom you want to generalize your results to. Again this does not apply to qualitative research as these studies do not want to generalize results to a wider population.
5. Clearly define your sample with inclusion and exclusion criteria. If you have a complicated sample, such as a cluster sample you also need to define the “clusters” with inclusion and exclusion criteria.
6. Sampling: Please consider different sampling strategies and read on sampling and sample size.
NOTE: Sampling is usually different in qualitative studies which want an information-rich sample. In contrast, quantitative studies usually want a representative sample using probability sampling.
7. Sample size: In a quantitative study the required sample size is directly linked to the falsifiable research hypothesis. Depending on the research hypothesis DIFFERENT formulas apply. The simplest applies when estimating the prevalence of a disease using a cross-sectional study However, formulas change if you want to compare groups of people, and again if your sampling design is complex (e.g. cluster sampling). Many sample size calculators are available online. Please ask us if in doubt!
Qualitative studies sample until the data reaches “saturation”! So the sample size in qualitative studies cannot be calculated beforehand.
8. Make measurements of the exposure and the outcome and potential confounders. Clearly describe and define your main variables in your study.
9. Plan how to analyze the results. Please read
NOTE: The statistical analysis of the study has to be linked back to your research hypothesis. How will you answer your research hypothesis with statistical techniques? Similarly for qualitative studies – the analysis has to be linked back to the research question. Although, qualitative studies do not apply statistics to reach conclusions.
10. Avoid bias in sampling, measurement and confounding. Please read The topic about bias is important when planning a study and also links to the limitations of the study in your research proposal where we expect an honest account of expected bias.
11. Consider the time frame and ethics.
NOTE: The time frame is about the feasibility of your study. Please consider the required sample size and how long it will take you to recruit the number of people.
Ethics should also reflect on the particular situation of the study and how participants might be affected by the study and how negative effects could be minimized.
12. Consider the potential limitations of the study. Consider selection bias (whether your sample represents your target population); information bias (whether the information on exposure and outcome are valid); and confounding (whether results might be distorted by other variables). Please also comment on how to minimise these biases in your study. Please read
NOTE: Qualitative studies are not so much worried about bias – but they need to show that their results can be corroborated with other findings. This is called “triangulation”. Please read the MSF document carefully if you are proposing a qualitative study.
RESEARCH SKILLS PROGRAM PART 2
The second part of the RSP provides you with resources that introduce quantitative and qualitative research methodology. Please use the RSP Part 2 as a support for developing your research proposal.
-
Consulter
This folder contains resources about research question and research hypothesis - resources to consider at when you begin writing of your research proposal (assignment 2).
-
Consulter
This folder contains resources about quantitative and qualitative study designs and also a resource about choosing the right design for your research question.
Many people choose quantitative research study designs, but don't forget that qualitative study designs prepared by MSF in the folder. Please consider whether a qualitative study is more appropriate than a quantitative study when you are picking your study design. Qualitative research is particularly relevant when you are attempting to gain information from subjects who are not able to read and write well, or are from a marginalised population whose voices are seldom heard. Qualitative research can provide answer to many relevant questions in public health practice, especially regarding implementation… So often the what to do is quite clear, but the why and how much less so
The WHO book "Basic Epidemiology" by Bonita, Beaglehole and Kjellstroem (download pdf here) advises on basic quantitative study designs.
-
This folder contains resources about sampling and sample size calculation for a very simple example. A good article specifically about sampling can be found at: http://www.oandp.org/jpo/library/1995_03_105.asp.
-
This folder contains resources about statistical analysis for quantitative studies.
-
This folder contains resources about bias in quantitative studies.
-
Consulter
Please work through the exercises.
-
Écrire des messages de forum : 1
This forum is for reflecting on what you have learned when working through the resources provided and when starting with your project proposal.
You might want to reflect on:
Creating a research hypothesis
The study design
Sampling techniques and sample size
Statistics for qualitative studies - if appropriate
Bias
What you learned from the exercises
-
-
PREPARE A STUDY PROTOCOL FOR A RESEARCH PROJECT
Here are the major headings you need to consider in designing a research project.
1. Pick your study design, relevant to the question. The design could use quantitative or qualitative methods, alone or in combination.
2. Then select your materials including sample to study, with clear inclusion and exclusion criteria and Sample size
3. Make the relevant measurements including of the exposure, outcome and potential confounders. Clearly describe and define your main variables in your study.
4. Plan the analysis of the results.
5. Consider potential challenges in the above whether during sampling, measurement and analysis, and how you could overcome them.
6. Consider the time frame and ethics.
-
Écrire des messages de forum : 1
Reflect on the following aspects of your research study:
The study design.
- Have you stated a study hypothesis and/or research question(s)?
- Is the design chosen clear, and is it appropriate for the study hypothesis or research questions?
- Have you thought about the choice of design against other possible options?
Target population and study samples.
- Are the target population and study samples consistent with the study hypothesis / research question(s)?
- Are the population and study samples adequately described and thought through.
- Are inclusion and exclusion criteria defined?
Sample size and sampling process.
- Has a sample size calculation been conducted?
- Is the sampling process clear and feasible?
Study methods and data collection.
- Are the methods and data collection clear and feasible?
- Is the measurement of the study and outcome factors clear and valid?
- Has measurement of potential confounding factors been considered?
- Have details of the analysis been considered?
Timeframe.
- Have major milestones been identified?
- Are the sequence and timing logical and appropriate?
Ethical considerations.
- Has institutional ethical approval been described?
- Do you plan to obtain informed consent?
- Has anonymity, confidentiality and privacy been considered?
- Is data handling/storage appropriate?
- Are there other ethical issues?
Limitations of study including potential bias.
- Have the limitations been considered?
- Might there be any biases, including selection, information and confounding?
- Have processes to reduce bias been identified (if appropriate)?
-
-
This is a long course, with many components. Although this does not cover all the components, you can gain a certificate of completion if you have posted to each of the Reflection forums.
