After you set up this table, set up some more criteria that will help you to narrow down your search and retrieval of literature that you will use for your evidence map. For example, in our study we have decided that we will include only the following:
- All studied published since 2019
- Every study must only be published in English language or an English translation must be available
- All studies must be peer reviewed and standard publications (this means we will exclude letters to the editor, editorials, popular non-peer reviewed magazine articles, books, book chapter, monographs that are not primary studies, etc).
- All studies must be freely available in full text format
You can create further restrictions or you can relax restrictions as the case may be. The less restrictions you will have , the more wide you can cast your "net" for fishing articles, and the more number of articles and primary sources you will use.
Step 2: Search literature
After you complete the first step of framing a pico formatted question, you will need to formulate search terms. The search terms can be framed in plain English, and you will also need to access the Pubmed MeSH vocabulary (link:
https://www.ncbi.nlm.nih.gov/mesh) to identify what might be correct term to search. The search terms can be anywhere in the following areas of a paper or a monograph or a research report:
- Terms can be found in the title, abstract, or text body
- Terms can be combined using a logic referred to as Boolean logic (if you do not know what is meant by Boolean logic, make sure to read this article: https://www.lotame.com/what-is-boolean-logic/)
- You will need to use synonyms of the terms. For example, instead of heart disease, you can use "acute myocardial infarction", or "AMI", or cardiac illness and other terms. It is a good idea to spend time to identify the various synonyms that you are likely to come across.
After you set up the terms, think of which databases you will search. If you are working with environmental health related data sets, I recommend you search at least:
- Google Scholar:
- Pubmed/Medline
- University Library (University of Canterbury Library databases for example)
- Microsoft Academic
- Agricola
- Web of Science
- Other databases that are relevant to your discipline
- Campbell Collaboration database
- Cochrane Collaboration database
For example, for our study, we used the following search terms and we only searched Google Scholar. But in your case, you may want to search more than one database and want to include more search terms in various combinations. Note that the boolean operators "AND", "NOT", and "OR" when they are used, are used in capital letters or upper case letters.
These are our search terms
"air pollution" OR "criteria pollutants" OR "PM*"
AND "heart disease" OR "acute myocardial infarction" OR chronic heart disease"
Step 3: Read the abstract and select studies for reviewusing the ROSES form
Once you have obtained a set of titles and abstracts based on your literature database search, read the titles and abstracts of every study one by one. When you do so, read them with your inclusion and exclusion list in hand. Exclude the studies one by one if they do not meet the criteria of focusing on the topic of your interest, or if they have wrong population, of if they have wrong intervention or exposure, and so on. Make a note of how many studies you originally selected and how many studies you ended up in your data set that you will analyse.
When you have identified the studies, download or obtain their full texts and read the full texts of the papers. Then, read their reference lists and identify any potential study that will be useful. Obtain the title and abstract of that study and check if that paper can be included. If so, then obtain the full text and read. This is an iterative process, and as you do so, keep a tally of how many studies you started with, how many studies were rejected based on the various grounds of rejection, and how many you finally ended up with. For environmental health, use the ROSES (Reporting Standard for Systematic Evidence Synthesis) Protocol (URL:
https://www.roses-reporting.com/systematic-review-protocols) . Download the spreadsheet from the above site and fill in the ROSES checklist and the flowchart maker.
In our study, we did not exactly follow this sequence as we identified 10 studies published since 2019 and we decided to use all of them. We also did not read the reference lists of all the studies, and hence we did not exactly follow the process we outlined above.
Step 4: Create a data extraction form
Once you have settled on the exact number of studies, create a data extraction form. You can use a spreadsheet for this purpose, typically something like an Excel spreadsheet or a Google Sheets. The spreadsheet or the sheets app will enable you to create forms that you can distribute to research students or others such as you colleagues to enter data there. Or you can directly use a spreadsheet to enter data.
What elements will go into the form is subjective but below we recommend a minimum set of variables that must go into the form.
For this project we have the following items to be extracted from each study:
- study_id: based on the study outcome, last name of the first author, and the year
- author: name the first or corresponding author
- outcome: specify the outcome studied
- population: specify what population was studied
- exposure: specify and describe the particular exposure studied
- comparison group: what was compared with respect to the exposure (the comparator must be described with respect to the exposure variable)
- year of the study: specify which year the study was completed
- finding: summary finding (one word)
- study design: what epidemiological study design was used to study the association between exposure to air pollution and heart disease
After this step, we obtained the following spreadsheet