Access provided by
London Metropolitan University
• To appreciate the use of Microsoft Excel for systematic reviews
• To increase your awareness of all the steps of a systematic review, including importing references and documents, plus documenting the whole process, including intermediate results
• To learn how to create transparent and complete reports for systematic reviews
Background Systematic reviews are complex and involve time-consuming, deep research in the academic literature to search, extract data, assess their quality and report the results. Digital tools and software have been developed to simplify different phases of this process but some of these are not free to use. Microsoft Excel is typically accessible to a great many researchers free of charge, so using it involves no further costs.
Aim To explain how to use Microsoft Excel to create transparent and complete reports for systematic reviews.
Discussion The author’s method includes six steps: downloading the references, preparing worksheets, removing any duplicate references, screening the references by title and abstract, screening the full text of references, and listing the articles for inclusion in the review.
Conclusion The Excel method is efficient and free and can produce transparent and complete reports of systematic reviews. It is a valid alternative to the systematic reviews produced by advanced tools and software.
Implications for practice The documents produced by this method are a good source for the direct production of scientific texts.
Nurse Researcher. doi: 10.7748/nr.2023.e1866
Peer reviewThis article has been subject to external double-blind peer review and checked for plagiarism using automated software
Correspondence Conflict of interestNone declared
Godino L (2023) How to structure Microsoft Excel documents for systematic reviews. Nurse Researcher. doi: 10.7748/nr.2023.e1866
AcknowledgementThe author would like to thank her colleagues and students at the University of Bologna for their interest in this method, as well as the people who encouraged her to write this article
Published online: 01 March 2023
This article introduces a cost-effective way to use the spreadsheet program Microsoft Excel to create a transparent and complete report of a systematic review that provides sufficient documentation for the direct production of scientific texts. The aim is to establish a more rigorous and controlled process for conducting systematic reviews that does not introduce the biases typical of traditional reviews.
Colleagues, master’s students and I have used the Excel method for several years, during which we have refined, honed and applied it in different reviews (Godino and Skirton 2012, Godino et al 2013, 2016, Bertonazzi et al 2022). It is now time to make it available to all those struggling with the enormous number of documents that systematic reviews identify.
A systematic review is a primary research method that enables researchers to gather, evaluate and analyse all the articles available about a given research question (Biolchini et al 2005). It is ‘a replicable, scientific and transparent process… that aims to minimise bias through exhaustive literature searches of published and unpublished studies and by providing an audit trail of the reviewer’s decisions, procedures and conclusions’ (Tranfield et al 2003). Systematic reviews are commonly used in healthcare as baselines for developing clinical practice guidelines, and for professionals to keep up to date with developments in their fields. They are also fundamental for stakeholders and healthcare providers, who would otherwise face a considerable amount of research on which to base their decisions.
Researchers who conduct systematic reviews are more likely to obtain unbiased and complete descriptions of the literature than if they conduct traditional reviews of the literature (Page et al 2021). The latter might introduce biases at different points, from the formulation of the research question to the collection, evaluation, analysis, interpretation, summarisation and presentation of the data. The value of a systematic review depends on what was done, what was found and the clarity of reporting, which should be transparent and complete to enable other researchers to repeat or update the review.
A systematic review does not consist simply of rearranging of what is already known – it is a research method, with a rigorous methodology that includes a strict, well-defined sequence of steps. A few guidelines have been developed over the past decade with the aim of supporting researchers in preparing transparent reports from systematic reviews. One such guideline – Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Moher et al 2009, Page et al 2021) – has been widely endorsed and adopted.
PRISMA consists of four steps:
1. The researcher identifies articles relevant to the study.
2. They screen the articles using predetermined criteria.
3. They methodically classify the articles using predetermined codes and themes.
4. They decide which to include in the systematic review.
This process can be time-consuming, given the rigour it requires and the lack of tools or software available to support and speed up the different steps (O’Connor et al 2019). Van der Mierden et al (2019) evaluated 16 tools used for systematic reviews and found none of them supported the complete process; DistillerSR, EPPI-Reviewer, Covidence and SWIFT Active Screener were found to be the best tools but were not free to use.
However, Microsoft Excel is available and is used by many professionals around the world as part of Microsoft Office. Excel’s reasonable price and wide popularity means a simple method for systematically selecting and reporting articles based on basic Excel functions can be suitable for researchers and students.
This method requires only Microsoft Excel and uses only a few of its common functions. It assists in all the steps of a systematic review, including importing references and including documents, as well as in documenting the whole process, including intermediate results. It helps to:
• Report the number of recovered articles.
• Exclude articles before screening.
• Report the articles after screening, via their titles or their titles and abstracts.
• Retrieve articles for meticulous evaluation.
• Identify potentially eligible articles that were unrecoverable.
• Find the articles that did not meet the inclusion criteria and determine the main reasons for their exclusion.
• Enumerate the number of articles included in the review.
The method consists of four phases and six steps (Figure 1).
First download all the references you found from the databases you used for your literature search. You can usually do this using functions such as ‘Export’ and ‘Save citations’. Choose either Excel or CSV as the format for the download.
Save each file in the same folder on your device. Rename it so that it includes the database from which you downloaded it and the date you downloaded it – for example, ‘Pubmed_22.05.22.xlsx’.
Create an Excel workbook that has worksheets with the following names or similar: ‘All references’; ‘All references with duplicates’; ‘All references without duplicates’; ‘Full-text screening’; and ‘Articles included’. The order in which you arrange the worksheets does not matter – all the references retrieved are stored in a single workbook and you can analyse the whole systematic review process with this method. It can therefore help to answer questions such as ‘Why did I not include this article?’ by checking your search, for example.
Add the following headings to the worksheets ‘All references’ and ‘All references with duplicates’: Database, Title, Authors, Abstract, Year of publication, Language, Journal, Document Type and Digital Object Identifier. Now open each file you downloaded in Excel. Select all the references, copy them then paste them into the ‘All references retrieved’ and ‘All references with duplicates’ worksheets.
Create another worksheet named after the database from which you downloaded the file – for example, ‘PsychoDB’. Add the headings to the worksheet, then copy and paste the references into it as well, making sure that the columns are in the same order as the headings – rearrange them to match if necessary; for the database column, put the name of the database in every cell.
Once you have imported all the references into the workbook and ordered them, you must check to see if there any duplicates and remove them.
Go to the worksheet ‘All references with duplicates’. Select all the data then go to Excel’s ‘Data’ tab and click the ‘Sort’ icon. In the new ‘Sort’ window that appears, click the ‘+’ button so there are three levels – ‘Title’, ‘Authors’ and ‘Database’ – and choose to order each ‘A to Z’ then click OK.
Add a new column called ‘Duplicates’ to the left of the ‘Database’ column in the worksheet. Select all the references then select the ‘Home’ tab. Click the ‘Conditional Formatting’ icon and select ‘Highlight Cells Rules > Duplicate Values’ from the dropdown menu.
In the ‘New formatting rule’ window, use the ‘Format with:’ menu to pick how you want Excel to highlight duplicate values – for example, ‘Green Fill with Dark Green Text’. Then click OK.
Select each highlighted reference. Enter 1 in the ‘Duplicates’ cell of each of them. To calculate how many duplicate references there are, use the formula ‘=SUM(…)’ in an empty cell below the list of references to add all the values of the cells in the ‘Duplicates’ column.
Now, select all the references and headings, copy them, then paste them into the worksheet ‘All references without duplicates’. Sort the references again, this time by ‘Duplicated’ in ‘descending order’. Select and delete all the references highlighted in grey, which will be in one block.
You will now have a list of all the references that contains no duplicates.
Screening is the most labour-intensive part of a systematic review, since you must read the title and abstract of every article to remove references that are irrelevant or inappropriate for the research question.
First, add a new column called ‘To read full-text’ to the left of the ‘Database’ column in the worksheet ‘All references without duplicates’. Then create new columns to the left of ‘To read full-text’ for each of your inclusion and exclusion criteria.
Now read the title and abstract of each reference. If you decide the reference meets a criterion, enter 1 in the cell corresponding to that criterion. Do not continue evaluating the remaining criteria, once you determine a reference is ineligible for inclusion in the review. If you decide the article is suitable for inclusion after checking all the inclusion and exclusion criteria, enter 1 in the cell ‘To read full-text’.
When you have screened all the references, create a new cell with the formula ‘=SUM(…)’ for all the exclusion criteria, as well as for ‘To read full-text’. This will give you the number of excluded references and the number whose full text you need to read The result should look similar to Figure 2.
Full-text screening is conducted similarly. Select everything in the ‘All references with no duplicates’ worksheet and paste it into the ‘Full-text screening’ worksheet. Sort the worksheet by ‘To read full-text’ in ‘descending order’, then select and delete all the references that do not have a 1 in that the column to obtain the list of all the articles to read. Now add a new column, ‘To include’, before the ‘Database’ column.
Read the full text of each referenced article. If you decide the article is eligible for inclusion in the systematic review synthesis, enter 1 in the ‘To include’ column; you should also enter 1, if you cannot retrieve the full text of an article.
Create a new cell under the final reference with the formula ‘=SUM(…)’ for values of the ‘To read full-text’ column, to count how many references you included; you can then calculate how many you excluded by subtracting the result from the total number of references you downloaded.
You should now be able to easily show which references you retrieved and the primary reasons why you excluded a reference, at the ‘Eligibility’ stage of the PRISMA flowchart.
Lastly, copy all the headings and references in the worksheet ‘Full-text screening’ and paste them into the worksheet ‘Articles included’. Sort the references by ‘To include’ in ‘descending order’.
All the references suitable for inclusion in the review will be in one block; select and delete the remaining references to create a list of only the suitable articles (Figure 3)
You can duplicate the ‘Articles included’ worksheet to use it when assessing the quality and the synthesis of the references included in the review. You should also report how many references you included, at the ‘Included’ stage of the PRISMA flowchart.
A fundamental step in a systematic review is describing the results of your research and selection process to ensure readers understand the inclusion criteria you chose. For example, reporting the number of references retrieved from literature databases will be useful for repeating the search.
PRISMA flowcharts provide a record of the flow model used for the process, broken down by source (Page et al 2021). Using the Excel method makes the PRISMA flowchart simple to adopt, as the worksheets show the number of references retrieved from databases, the numbers from the screening phases, and the references to be included in the qualitative and quantitative syntheses.
Your final flowcharts should look like Figure 4 if you are using PRISMA 2020 or Figure 5 if you are using PRISMA 2009.
All articles selected for inclusion in a systematic review must be subject to independent quality assessment. Several tools are available to assist researchers in this, but the Excel method uses criteria validated by Kmet et al (2004) as these provide a comprehensive manual for assessing the quality of quantitative and qualitative studies.
Kmet et al’s (2004) method involves evaluating specific characteristics of an article pertaining to its methodology and how its results are reported. Ten characteristics are evaluated for qualitative studies, 14 characteristics for quantitative studies. The study is awarded zero points if it does not address a characteristic, one point if it partially addresses it and two points if it fully addresses it. If a characteristic is inappropriate for a quantitative study’s chosen design, that characteristic should instead be marked ‘n/a’ (not applicable). The marks can be recorded and the corresponding quality scores obtained using an appropriate Excel spreadsheet.
The total score is calculated by multiplying the number of fully addressed characteristics by 2 then adding the number of partially addressed characteristics. For a quantitative study, the total possible score is determined by multiplying the number of characteristics assigned ‘n/a’ by 2 and subtracting the result from 28; ‘n/a’ cannot be applied to any characteristic in a qualitative study, so the total possible score is 20. The ratio of the total score to the total possible score is then determined as a percentage to indicate the study’s quality.
Total scores and percentages should be reported in the systematic review. Kmet et al (2004) did not impose a minimum score for a study to be included in a review, although the authors suggested 60% was a reasonable cut-off point.
Two independent researchers should assess the quality of the review. Each researcher should each have a dedicated column in the worksheet for the assessment. They can discuss any disagreement about scores until they reach a consensus.
The Excel method is intended to streamline the screening and extraction processes of a review. However, not considering its limitations could lead to an increased workload rather than a simpler one.
Overall, it has the following strengths:
• It enables researchers to screen and review the whole process.
• It is inexpensive.
• It is simple to navigate.
• It ensures researchers accurately report numbers.
• It has innovative data extraction features.
Its weaknesses are:
Systematic review is a research methodology that synthesises and integrates the results of multiple empirical studies of a specific area of interest. However, conducting a review is not easy, with most researchers commonly encountering problems ranging from importing references to including articles. This has highlighted the need to develop systematic implementation methodologies, such as the Excel method described in this article.
Effectiveness of a neutropenic sepsis clinical pathway
This article reports the findings of a re-audit across a...
Identifying patients at risk of neutropenic sepsis
An audit of mortality figures for patients undergoing...
The biology of cancer
Cancer research is moving fast. Understanding of the biology...
Compliance with swift communication of diagnosis to GPs
Until this year the National Cancer Peer Review Programme...
Improvement of chemotherapy practice and quality of care
There has been increased pressure on chemotherapy day units...