Systematic Review

Narrative Review

A narrative review covers a broad range of topics and is conducted without a standardized methodology. Search strategies and the degree of comprehensiveness may vary, as no fixed protocol is followed. 

Objective: To obtain an initial overview of the state of research and develop possible hypotheses.
Duration: approx. 1 to 12 months

Rapid Review 

A rapid review applies key elements of systematic reviews in a simplified form. 

Objective: To provide a summary of the current state of the literature within a very limited timeframe, in order to make well-informed decisions – such as regarding treatment options –quickly.
Duration: approx. 1 to 3 months

Scoping Review 

A scoping review systematically collects and categorizes the current state of research on a broad topic or a series of research questions. This process may include a critical appraisal of the literature. Unlike systematic reviews, however, this is done without an evaluation based on formal quality criteria and without a targeted synthesis of the results.

Objective: To identify research gaps and future research needs.
Duration: approx. 6 months

Umbrella Review

An umbrella review synthesizes the results of multiple systematic reviews and meta-analyses that address similar research questions. Unlike a systematic review, the research question is often much broader in scope. 

Objective: To gain a comprehensive overview of existing systematic reviews and meta-analyses.
Duration: approx. 3 to 6 months

Meta-Analysis

A meta-analysis is a statistical technique used to quantitatively evaluate the results of multiple stud-ies. It is often based on a systematic review.

Objective: To combine the results of multiple studies statistically in order to obtain reliable effect sizes.
Duration: depending on the number of studies identified in the systematic review.

Formulating an appropriate search query is the key step in preparing for your search. 

1. Formulate a research idea.
2. Conduct “scoping searches” (= preliminary exploratory searches in a few databases to get an idea of the scope of current research on the research idea).
3. Develop an initial research question: Describe the most important aspects in your own words and derive a research question from them. 
4. Refine the research question using a question framework, such as PICO:

Additional potential question frameworks can be found in this Rapid Review von BMJ Glob Health. 

Once the research question has been formulated, the inclusion and exclusion criteria are defined to determine study eligibility in the systematic review—even before the literature search begins. 

Possible and very common inclusion and exclusion criteria:

• Publication type, e.g., systematic reviews, meta-analyses, journals with a specific ranking, peer-reviewed articles
• Study design, e.g., randomized controlled trials
• Type of population, e.g., age, gender, pre-existing conditions
• Type of intervention/control group – Specific features of the intervention
• Type of measurement, e.g., determination of laboratory values using a prescribed method
• Duration/temporal context, e.g., measurement taken five years after diagnosis, publication period

A protocol establishes clear, structured guidelines for the systematic review. When drafting a protocol, you can use the PRISMA-P Checklist as a guide. To avoid duplication of effort and achieve the highest possible level of transparency and reusability, the protocol should be registered. The ideal time to do this is as soon as the research question has been formulated, but data extraction has not yet begun.

Possible registration platforms: 

The selection of a suitable database or platform depends on the specific research question and the respective field of study. The Database Information System (DBIS) offers an initial guide to suitable databases.

In the health sector, you can also use RefHunter. This tool helps you, among other things, compare the functionalities and search options of various databases. Additionally, we recommend searching study registries to obtain information on currently ongoing studies (e.g., DRKS, ClinicalTrials.gov, or WHO ICTRP). 

Additional search options:

• Searching repositories
• Searching in specific journals 
• Manual searches and gray literature, e.g., preprint servers, dissertations and Theses, NGO and foundation reports, specialized academic libraries, government and agency sources

The success of a systematic review depends to a large extent on selecting the right search terms. The goal is to generate a hit list that is as comprehensive as possible while remaining precise. We recommend combining keywords and subject headings.

Keywords

• Are derived from titles, abstracts, and author information
• Reflect current terminology

Subject heading/Descriptor / Thesauri from databases

• Ensure consistency despite different spellings
• Examples: MeSH (PubMed/Medline), Emtree (Embase), AGROVOC (Agris

Selection of more helpful tools:

To construct a search syntax, you need Boolean operators:

• OR → expands the search → synonyms, related terms, abbreviations
• AND → narrows the search → thematic intersection
• NOT → excludes → use with caution because of bias risk

Step-by-step guide

• Break down your question into aspects. 
• For each aspect, identify as many synonyms, related terms, and abbreviations as possible. 
• Start by searching for these terms individually in the database’s search platform. This will give you an initial idea of the expected number of results and allow you to identify any typos right away. 
• Combine the synonyms of an aspect using OR to form aspect bundles. (Aspect A1) OR (Aspect A2) OR (Aspect A3)
• Combine the aspect bundles using AND. 
• Review the results.

*Please note: Syntax depends on the search platform.

Familiarize yourself in advance with the search mechanisms and functions of the respective data-base. You can find the syntax and, as a rule, examples in the search platform’s help section.

The search process is usually iterative: develop → test → refine. If the number of hits does not meet your expectations, you can take the following steps, for example:

Reduce the number of results

• Use subheadings
• In PubMed: Restrict to “Major Topic”
• Use filters/limits (e.g., publication type, age, human/animal, time period)
• Phrase search
• Use operators such as AND or proximity operators
• Add an aspect

Expand the number of hits

• Select a higher-level term
• Include subterms in the search (e.g., MeSH: explode)
• Include as many relevant search terms as possible (MeSH + keywords, synonyms, terms from search results, other languages)
• Consult encyclopedic sources like Wikipedia, or AI-based tools for additional terminology, use truncation and masking
• Include abbreviations and alternative spellings
• Omit aspect

In the course Literature Research Advanced – Finding Relevant Scientific Papers in Databases, we will discuss how to conduct efficient literature research using databases. The starting point is the targeted selection of relevant databases and a well-structured search strategy that optimizes your search and delivers precise results. Last but not least, we will explore ways to access academic papers. A solid foundation for successfully navigating research in the database universe and systematic review!

Go to Course