• Systematic review (completion time 15–30 months)
  Systematic searching, evaluation and summarisation, formal quality assessment procedure.

• Rapid review (completion time 1–3 months)
  Same as systematic review, but shorter, simplified and with reduced methodology.

• Scoping review (completion time 6 months)
  Evaluating the size and scope of available research literature and identifying gaps and research needs

• Umbrella review (completion time 3-6 months)
  Summarising the results of several reviews

Tip

With the RightReview tool, you can find the review type that is right for you.

Before starting your SR, you should find out whether there already is an SR on the topic or whether one is planned. To do this, we recommend searching PROSPERO (International Prospective Register of Systematic Reviews) and the Cochrane Library (medicine).

Formulate a precise research question

• searchable
• neither too narrow nor too broad
• focused on available and accessible information

Develop a research protocol

• In the introduction of your research protocol you should introduce the topic, state your specific research question and explain the relevance of the work.
• In the main section, you explain the current state of research in which your work is embedded. Against this background, you develop your research proposal and the analytical framework of your work. This should be followed by a work schedule in which you explain the essential steps of your project.
• Conclude your research protocol with a bibliography.

To search for literature, the research question is divided into individual search components to bring it into a format that can be searched in databases.

In medicine and related subjects, the PICO framework (patient/population, intervention, comparison, outcome – additions: time, context/setting, study design) has proven to be effective. In most cases, inclusion and exclusion criteria are also defined (age, concomitant medication, pre-existing conditions).

Note: As an alternative to the PICO scheme, frameworks such as SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research Type) or SPICE (Setting, Perspective, Intervention, Comparison, Evaluation) are suitable for qualitative or theory-based questions. Additional information can be found in the comparison by the Australian National University Library and the list of numerous frameworks by the University of Maryland.

Example of a research question in medicine:

What is the efficacy and safety of GLP-1 receptor agonists compared to insulin therapy in adult patients with type 2 diabetes mellitus in terms of blood glucose control, weight change and cardiovascular events?

• P (Population): Adult patients with type 2 diabetes mellitus
• I (Intervention): Treatment with GLP-1 receptor agonists (e.g. liraglutide, semaglutide)
• C (Comparison): Conventional insulin therapy (e.g. basal insulin, intensive insulin therapy)
• O (Outcome): HbA1c value (blood sugar control), body weight, cardiovascular events (e.g. myocardial infarction, stroke), side effects

Example of a research question in clinical psychology:

Is cognitive behavioural therapy (CBT) more effective than psychodynamic therapy in treating depression in adults?

• P (Population): Adults with clinically diagnosed major depression (according to DSM-5)
• I (Intervention): Cognitive behavioural therapy (CBT)
• C (Comparison): Psychodynamic psychotherapy
• O (Outcome): Reduction in depressive symptoms (e.g. measured using BDI-II or PHQ-9)

• When conducting systematic literature searches, select databases that are relevant to your topic and justify your selection.
• The quality and relevance of the databases are more important than their number.
• Databases may overlap in terms of content, but they have different areas of focus.
• A database can be provided by various providers with different search functions (e.g. MEDLINE via PubMed, EBSCO, Ovid, ProQuest, Web of Science).
• Meta search engines allow simultaneous searches in multiple databases via a single interface.
  Disadvantage: The search method in the individual databases is often not transparent. Meta search engines are suitable for initial orientation or as a supplementary search tool.

Tip

• Specialised databases at RUB
• DBIS - Database Information System
• RefHunter

• Keywords are central terms within references (e.g. in the title, abstract or full text).
• When searching for keywords, make sure that they correspond to the search language of the respective database (e.g. use English search terms for an English language search interface).
• Consider synonyms, abbreviations, translations and other spellings (e.g. British/American English).
• The keywords you found in this step can be used in all selected databases.

Tip
Use the text analysis tool PubReMiner

PubReMiner is a tool that provides statistical analysis of PubMed search results. As a simple, text-based tool for creating search queries, PubReMiner (PRM) evaluates the information from PubMed results and displays the results in the form of frequency tables – sorted by journal, author or keyword. For example, you can enter the PubMed IDs of the most relevant hits and identify words that occur most frequently in the titles and abstracts of the articles found. You can easily adjust your search query and send it to PubMed for a rerun.

• Subject terms are words that describe the content of a document.
• Databases often contain standardised subject term lists or thesauri (tree structures with main and sub-terms, e.g. the MeSH terms in PubMed).
• Subject terms are database-specific, i.e. they must be identified separately for each database.
• When searching for subject terms, it is important to ensure that they correspond to the search language of the respective database (e.g. use English search terms for an English language search interface).
• Subject term searches enable very targeted searches with a higher proportion of relevant hits, as they can be used to find all hits indexed by the respective database on a topic (or all hits to which the respective subject term has been assigned).

When developing the search string, the individual search components of the query or the keywords and subject terms identified for it are combined using Boolean operators. The most important Boolean operators are:

• AND: all terms are included
  OR: one or both terms are included
  NOT: the term following NOT is not included

Use search techniques:
Include one or more of these techniques in the search string to expand or narrow the search, e.g.:

• Truncation and wildcards
  * often replaces any number of characters, ? often replaces exactly one character, $ often replaces zero or one character – Check the specific database Help section to find out which truncation/wildcard options to use: an asterisk / dollar sign / question mark / hash
• Phrase search (put the term in ‘’)
• Proximity operators (allow you to search for two or more search terms that occur within a certain distance of each other (e.g. NEAR/n)
• Use search commands for the specific search fields of the individual databases

Note 1
Please look at the search functions in each database, as they differ from one another.

Note 2: Truncation/wildcard search
Such a search does not necessarily yield more results than the same search without truncation/wildcards. This is because some databases do not apply multilingual search functions (such as stemming/lemmatisation, synonym mapping and spelling normalisation) to those searches. For example, a keyword search for ‘music theatre’ may return more results than a truncation search for ‘music theatre*’ because, without truncation, the database automatically searches for ‘music theater’ as well.
Performing a truncation or wildcard search does not necessarily improve the relevance ranking. In some cases, it can even have a negative effect on the relevance ranking, as some relevance factors, such as phrase match boosting and term weighting, are not applied to those searches.

Unfortunately, ONE search string does not work for all databases. Different databases use different search syntax, operators, and default search fields. When you are remapping your master search to your chosen databases, it is important to consider whether you will need to adapt any aspects of your search strategy such as phrase searches, truncation, wildcards, and proximity.

Use the database-specific options of your selected databases. The help pages or tutorials for the respective database may be useful here.

Example of a search string in the Title (TI) and Abstract (AB) search fields:

PubMed: exercis* [tiab]
Web of Science: TI=exercis* OR AB=exercis*

Tip 1
Database descriptions
You can find good descriptions of the search functions of over 100 databases at RefHunter.

Tip 2
Translation tools
The Systematic Review Accelerator tool allows you to customise the search string for many databases.
Other translation tools:

• Polyglot Search Translator
• LitSonar

Or use AI tools like ChatGPT or GPT@RUB: (PROMPT „Convert this search (your search string) into a search string appropriate for the Web of Science database")

Tip 3
Review
Review all automatically translated search strings!

To remove duplicates in a systematic review, use reference management software such as EndNote, Zotero or Mendeley. You can also use the free tool Deduplicator.

Tip

• Create subfolders for each database
• Create backup copies

An important aspect for the traceability of literature search is the careful documentation of the search process and the saving of the search string used. The PRISMA guidelines recommend to transparently disclose search queries for all databases used – including all Boolean operators, proximity operators, truncations and other search parameters used.

Tip

• Use the PRISMA checklist
  The PRISMA checklist (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) is a 27-item checklist used to improve the quality and transparency of systematic reviews and meta-analyses. It covers all aspects of the report, including the title, abstract, introduction, methods, results, discussion and funding.
• Use the RefHunter search protocol

• Read other publications (SR)
• Use the PRISMA checklist
  The PRISMA checklist (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) is a 27-item checklist used to improve the quality and transparency of systematic reviews and meta-analyses. It covers all aspects of the report, including the title, abstract, introduction, methods, results, discussion and funding.
• Use editors
  - Create PRISMA Flowchart
  - Create a methodology report

Tip
Reference management software
assists you in preparing the literature you have found for use in your own academic work.

• Use the title and abstract to assess whether the publication in question should be included in the review.
• Then review and evaluate the full texts.
  Recommendation: Use software  (e.g. ASReview | Covidence | Nested | Systematic Review Accelerator) to select publications.
• Dual control principle
  Where possible, screening processes should be carried out by at least two people to avoid BIAS (distortions, e.g. incorrect exclusion of relevant publications, influencing inclusion or exclusion).

If you have any questions about scientific publishing in Open Access (funding opportunities, publication options, general advice), please contact the OA team at the University Library.