Databases, Search Platforms, and Search Resources in Systematic Reviews

Introduction

A systematic review depends on a transparent and reproducible process for identifying relevant evidence. An important but frequently misunderstood aspect of literature searching is the distinction between a database and the platform or interface used to search that database.

The distinction can be summarised simply:

A database determines what collection of records is being searched, whereas a platform determines how that collection is searched.

For example, MEDLINE is a bibliographic database produced by the U.S. National Library of Medicine (NLM), whereas Ovid is one of the platforms through which MEDLINE can be searched. Similarly, CINAHL is a bibliographic database, while EBSCOhost is the platform commonly used to access it.

The distinction matters because different platforms may use different search syntax, field codes, proximity operators, subject-heading functions, filters, and search-processing rules. Consequently, reporting only the database name may not provide sufficient information to reproduce a systematic-review search.

PRISMA-S, the extension of the PRISMA Statement for reporting literature searches in systematic reviews, emphasises transparent reporting of the information sources and search methods used in a review.[1] PRISMA 2020 similarly requires authors to specify the databases, registers, websites, organisations, reference lists, and other sources used to identify studies, together with the date on which each source was last searched.[2]

Learning objectives

By the end of this study note, the reader should be able to:

1. Distinguish between a bibliographic database and the platform or interface used to search it.
2. Differentiate bibliographic databases from scholarly search engines, publisher platforms, and broader search resources.
3. Recognise common database–platform combinations used in health and medical research.
4. Explain why the same database searched through different platforms may require different search strategies.
5. Report electronic information sources clearly and reproducibly in a systematic-review protocol or manuscript.

Database and platform: the fundamental distinction

What is a bibliographic database?

A bibliographic database is an organised collection of records describing scholarly publications. Depending on the database, a record may contain:

• article title;
• authors;
• journal title;
• publication year;
• abstract;
• author keywords;
• controlled subject headings;
• publication type;
• institutional affiliations;
• digital object identifier (DOI);
• database-specific identifiers; and
• other indexing information.

Examples of major bibliographic databases relevant to health research include:

• MEDLINE;
• Embase;
• CINAHL;
• APA PsycINFO;
• the Cochrane Central Register of Controlled Trials (CENTRAL);
• Scopus; and
• Web of Science Core Collection.

The database determines the underlying collection of records and how those records are indexed.

MEDLINE, for example, is the NLM’s principal bibliographic database and uses Medical Subject Headings (MeSH) to index its literature.[3] Embase is a separate biomedical database produced by Elsevier and uses the Emtree controlled vocabulary.[4] CINAHL focuses particularly on nursing and allied-health literature.[5]

What is a search platform?

A search platform, sometimes called a search interface or provider interface, is the software environment through which a database is searched.

The platform controls functions such as:

• how search terms are entered;
• Boolean operators;
• truncation and wildcard syntax;
• proximity searching;
• field searching;
• phrase searching;
• subject-heading selection;
• exploding subject headings;
• search-history functions;
• filters and limits;
• record display;
• exporting results; and
• saving or rerunning searches.

Examples of search platforms include:

• Ovid;
• EBSCOhost;
• Embase.com;
• the Cochrane Library; and
• Web of Science.

A useful conceptual model is therefore:

Database → content

Platform → search interface

For example:

MEDLINE → Ovid

means:

MEDLINE is the database, and Ovid is the platform used to search it.

Similarly:

CINAHL → EBSCOhost

means:

CINAHL is the database, and EBSCOhost is the platform used to search it.

A useful analogy

The distinction can be understood by thinking about a physical library.

Imagine that a library contains a large collection of books.

• The collection of books is analogous to the database.
• The library catalogue used to find those books is analogous to the search platform.

Changing the catalogue system does not necessarily change the underlying collection, but it may change how users search for items within that collection.

In evidence synthesis, this distinction is particularly important because search strategies must often be adapted to the search platform being used.

Common databases and their search platforms

The following table summarises several important resources used in medical and health research.

The table demonstrates why terms such as MEDLINE, Ovid, PubMed, Google Scholar, and ScienceDirect should not automatically be grouped together and described simply as “databases”. They represent different types or levels of information resources.

MEDLINE and PubMed: an important distinction

MEDLINE and PubMed are often treated as interchangeable terms, but they are not identical.

MEDLINE

MEDLINE is a bibliographic database produced by the NLM. Articles selected for MEDLINE are indexed using MeSH and additional curated metadata.[3]

MEDLINE can be accessed through different search systems, including commercial platforms such as Ovid and EBSCOhost.

A systematic-review author might therefore report:

MEDLINE via Ovid

or:

MEDLINE via EBSCOhost

PubMed

PubMed is a broader NLM biomedical literature search resource. MEDLINE is its largest component, but PubMed also contains records from sources outside the MEDLINE subset.[6]

Therefore:

PubMed is not synonymous with MEDLINE.

When investigators search the whole PubMed resource, it is appropriate to report the source simply as:

PubMed

A phrase such as:

PubMed, including MEDLINE

may be useful when explaining the scope of the search to readers who may otherwise expect MEDLINE to be listed separately.

However, searching PubMed should not be described as exactly equivalent to searching MEDLINE through another platform.

Conceptually:

PubMed
│
├── MEDLINE records
│
└── Additional PubMed records

This also means that searching both MEDLINE via Ovid and PubMed may result in considerable overlap, although PubMed may contain additional records outside the MEDLINE component.

Embase and its platforms

Embase is a biomedical and pharmacological bibliographic database produced by Elsevier.[4]

It can be searched through platforms including:

• Embase.com; and
• Ovid.

Therefore:

Embase via Ovid

and:

Embase.com

should not be interpreted as two completely independent bibliographic databases.

They represent access to Embase through different search environments.

If investigators search Embase through both Ovid and Embase.com, they should recognise that considerable duplication is expected. Searching both interfaces should therefore have a clear methodological justification rather than being interpreted as searching two independent databases.

CINAHL and EBSCOhost

CINAHL, the Cumulative Index to Nursing and Allied Health Literature, is a bibliographic database focusing on nursing and allied-health literature.[5]

It is commonly accessed through EBSCOhost.

The appropriate reporting format is therefore:

CINAHL via EBSCOhost

Several CINAHL products exist, and institutions may subscribe to different versions. Where relevant, researchers should report the specific product that they searched, for example:

CINAHL Complete via EBSCOhost

rather than merely:

CINAHL

if the exact product is known.

CENTRAL and the Cochrane Library

The Cochrane Central Register of Controlled Trials (CENTRAL) is a database containing bibliographic reports of randomised and quasi-randomised controlled trials.[7]

CENTRAL is searched through the Cochrane Library.

Therefore:

CENTRAL = database

whereas:

Cochrane Library = the broader evidence resource and search environment through which CENTRAL is accessed

A clear reporting format is:

Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Library

Listing both:

• CENTRAL; and
• Cochrane Library

as though they were two independent databases would therefore be potentially misleading unless the investigator actually searched a separate Cochrane Library resource in addition to CENTRAL.

Scopus

Scopus is a multidisciplinary abstract and citation database produced by Elsevier.[8]

Unlike MEDLINE or Embase, where the database is commonly discussed separately from the search platform, Scopus is closely integrated with its own proprietary search environment.

It is therefore normally sufficient to report:

Scopus

There is generally no need to write:

Scopus via Elsevier

because this does not add meaningful information about the search interface.

Web of Science

The terminology surrounding Web of Science deserves similar attention.

Web of Science is the broader search platform, while the Web of Science Core Collection is a major citation database collection accessible through that platform.

Where the Core Collection has been searched, a more precise description is therefore:

Web of Science Core Collection

Authors may additionally specify the individual citation indexes searched when this information is relevant to reproducibility.

Google Scholar

Google Scholar differs fundamentally from databases such as MEDLINE, Embase, and CINAHL.

It is best regarded as a scholarly search engine rather than a conventional curated bibliographic database.

Google Scholar searches scholarly material from a wide variety of web-accessible sources. Its broad coverage can make it useful for:

• supplementary searching;
• identifying additional publications;
• citation searching;
• locating institutional or repository versions of articles; and
• searching for material that may not be retrieved through conventional bibliographic databases.

However, Google Scholar differs from conventional bibliographic databases in important respects. Its precise coverage and indexing processes are less transparent, and exhaustive reproducibility of very large search-result sets can be difficult.

For systematic reviews, Google Scholar is therefore commonly better described as a supplementary search source rather than being grouped with the core bibliographic databases.

ScienceDirect

ScienceDirect is also different from conventional bibliographic databases.

ScienceDirect is Elsevier’s full-text publisher platform, providing access primarily to journals and books published by Elsevier.[9]

This contrasts with Scopus, which is a citation and abstract database that indexes literature from many publishers.

The distinction can be summarised as:

ScienceDirect may be useful as a supplementary source for identifying or accessing relevant literature, but it should not automatically be treated as equivalent to comprehensive bibliographic databases such as MEDLINE or Embase.

Database, platform, search engine, and publisher platform

The different resources encountered during systematic reviewing can therefore be divided conceptually into several categories.

This classification explains why a statement such as:

“The databases searched were MEDLINE, Ovid, PubMed, Google Scholar and ScienceDirect.”

is conceptually imprecise.

These resources do not all belong to the same category.

Why the platform matters to the search strategy

The distinction between database and platform is not merely terminological. It has practical consequences for the search itself.

Search syntax differs between platforms

Different search interfaces may use different commands for:

• proximity searching;
• truncation;
• wildcards;
• title and abstract searching;
• phrase searching;
• subject-heading searching; and
• limiting results.

A search strategy developed for one platform should therefore not automatically be copied unchanged into another.

For example, a MEDLINE search written for Ovid may require modification before being implemented in another search interface.

The underlying research concepts may remain unchanged, but the search syntax must be translated into the language understood by the new platform.

Controlled vocabulary belongs primarily to the database

The database also influences the terminology used to index records.

For example:

A systematic-review search generally combines:

1. controlled vocabulary, when available; and
2. free-text terms, such as words appearing in titles and abstracts.

Therefore, translating a search from MEDLINE to Embase usually requires more than changing the platform syntax. MeSH terminology must also be translated into appropriate Emtree terminology.

A useful conceptual formula is:

Search strategy = research concepts + database-specific vocabulary + platform-specific syntax

Why identical concepts can produce different searches

Consider a review investigating an intervention for diabetes.

The underlying concept remains:

diabetes

However, the implementation may differ according to the database and platform.

For MEDLINE, researchers may use:

• relevant MeSH terms; and
• free-text terms such as diabet*.

For Embase, researchers may use:

• corresponding Emtree terms; and
• appropriate free-text terms.

For CINAHL, relevant CINAHL subject headings may be combined with free-text terms.

Thus, a high-quality search strategy is translated, rather than simply copied, from one database and platform to another.

Why this distinction matters for reproducibility

Suppose a systematic review reports:

MEDLINE was searched from inception to December 2025.

This identifies the database but does not fully describe how it was searched.

A reader does not know:

• which platform was used;
• which platform-specific syntax applied;
• how subject headings were processed;
• what field codes were used;
• whether particular platform filters were applied; or
• whether the published search strategy can be reproduced directly.

A more informative description would be:

MEDLINE via Ovid was searched from inception to 18 December 2025.

The complete Ovid MEDLINE search strategy should then be reported, usually in an appendix or supplementary material.

PRISMA-S provides specific guidance for reporting literature searches in systematic reviews and includes 16 reporting items addressing different components of the search process.[1]

Worked example: identifying the problem

Consider the following statement:

“The main databases to be searched are CENTRAL - Cochrane Central Register of Controlled Trials, CINAHL - Cumulative Index to Nursing and Allied Health Literature, CLIB - The Cochrane Library, Embase - Embase via Ovid, Embase.com, MEDLINE, PubMed, Scopus, Google Scholars, Science Direct.”

Several methodological and reporting problems are present.

Problem 1: databases and platforms are mixed together

The statement treats all resources as though they represent independent databases.

However:

• CENTRAL is a database;
• the Cochrane Library is the environment through which CENTRAL is searched;
• Embase is a database;
• Ovid and Embase.com are search platforms;
• Google Scholar is a scholarly search engine; and
• ScienceDirect is primarily a publisher-based full-text platform.

Problem 2: CENTRAL and the Cochrane Library are presented as separate databases

If the investigator intends to search CENTRAL through the Cochrane Library, these should be reported together:

Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Library

Problem 3: Embase via Ovid and Embase.com may represent duplicate searching

These represent different platforms for accessing Embase rather than two completely independent bibliographic databases.

Problem 4: MEDLINE and PubMed are treated as equivalent independent databases

MEDLINE is the largest component of PubMed, but PubMed also contains additional records.[6]

Searching both may be justified in some circumstances, but their substantial overlap should be recognised.

Problem 5: Google Scholar is incorrectly named

The correct name is:

Google Scholar

not:

Google Scholars

Problem 6: ScienceDirect is treated as equivalent to MEDLINE or Embase

ScienceDirect is primarily a publisher-based full-text platform and is more appropriately described separately when used for supplementary searching.

Example of a clearer search-source description

Suppose a systematic review intends to use:

1. CENTRAL via the Cochrane Library;
2. PubMed, including MEDLINE;
3. Scopus;
4. Google Scholar; and
5. ScienceDirect.

A clear description would be:

The electronic literature search will be conducted using the Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Library, PubMed (including MEDLINE), and Scopus as the primary electronic information sources. Search strategies will be adapted to the indexing structure and search functionality of each source using appropriate controlled vocabulary, where available, together with relevant free-text terms. Supplementary searching will be undertaken using Google Scholar and ScienceDirect to identify potentially relevant publications that may not be retrieved through the primary searches. Google Scholar will be used as a supplementary scholarly search engine, while ScienceDirect will be searched as an additional publisher-based full-text source. Reference lists of included studies and relevant review articles may also be examined to identify additional eligible studies.

This wording has several advantages:

• the core electronic sources are distinguished from supplementary searching;
• CENTRAL is correctly linked to the Cochrane Library;
• PubMed is not incorrectly described as being identical to MEDLINE;
• Google Scholar is not misclassified as a conventional bibliographic database; and
• ScienceDirect is appropriately identified as a supplementary publisher-based source.

How should databases be reported?

For systematic reviews, the preferred format is generally:

Database name + platform/provider + coverage or date range, where relevant + date searched

Examples include:

MEDLINE via Ovid, searched 18 August 2026.

Embase via Ovid, searched 18 August 2026.

CINAHL Complete via EBSCOhost, searched 18 August 2026.

Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Library, searched 18 August 2026.

For integrated resources such as Scopus or PubMed, reporting the resource itself is usually sufficient:

PubMed, searched 18 August 2026.

Scopus, searched 18 August 2026.

The full reproducible search strategy for each major database should additionally be documented.

Common mistakes and misconceptions

”Ovid is a database”

Not necessarily.

Ovid is principally a search platform that provides access to multiple databases.

A researcher should therefore normally specify:

MEDLINE via Ovid

rather than simply:

Ovid

”EBSCOhost is the same as CINAHL”

Incorrect.

EBSCOhost is the platform, whereas CINAHL is the database.

A researcher may use EBSCOhost to search several different databases.

”PubMed and MEDLINE are identical”

Incorrect.

MEDLINE is the largest component of PubMed, but PubMed also contains additional records.[6]

“CENTRAL and the Cochrane Library are two databases”

Not in the usual search context.

CENTRAL is the trial database. The Cochrane Library is the broader evidence resource through which CENTRAL is accessed.

”Searching Embase through Ovid and Embase.com means that two databases have been searched”

Incorrect.

These are different interfaces for Embase and are expected to have considerable overlap.

”Google Scholar is just another bibliographic database”

This is imprecise.

Google Scholar is better classified as a scholarly search engine.

”ScienceDirect and Scopus are similar because both are Elsevier products”

They serve different purposes.

Scopus is an abstract and citation database covering material from many publishers, whereas ScienceDirect is primarily a full-text platform containing Elsevier-published journal and book content.[8,9]

“A search strategy can simply be copied between databases”

Usually not.

The conceptual search can be preserved, but controlled vocabulary and platform-specific search syntax often need to be translated.

Practical decision points

When selecting and reporting a literature-search source, ask the following questions.

1. What exactly am I searching?

Identify whether the resource is:

• a bibliographic database;
• a citation database;
• a search engine;
• a trial register;
• a publisher platform; or
• another information source.

2. Through which platform am I searching it?

If the database can be searched through multiple platforms, identify the actual interface used.

For example:

MEDLINE via Ovid

rather than:

MEDLINE

when the platform is known and relevant.

3. Does the database use controlled vocabulary?

Examples include:

• MeSH for MEDLINE;
• Emtree for Embase; and
• CINAHL Subject Headings for CINAHL.

4. Does the platform use unique search syntax?

Check:

• proximity operators;
• truncation;
• wildcards;
• field codes;
• phrase-searching rules; and
• subject-heading commands.

5. Have I translated rather than copied the strategy?

The search should be adapted to the characteristics of each database and platform.

6. Have I recorded the search date?

Database contents change over time. The date on which each source was last searched is essential for reproducibility and updating.

7. Can another researcher reproduce the search?

Record and report:

• exact database or resource;
• platform;
• search date;
• search strategy;
• filters or limits;
• date coverage, where applicable; and
• any supplementary search methods.

Practical checklist for systematic-review authors

Before finalising the search-methods section, check that:

• Each major bibliographic database has been correctly identified.
• The search platform is specified where relevant.
• Databases and platforms are not incorrectly counted as separate information sources.
• PubMed is not described as being identical to MEDLINE.
• CENTRAL is correctly distinguished from the Cochrane Library.
• Google Scholar is identified as a scholarly search engine when used.
• Publisher platforms such as ScienceDirect are distinguished from multidisciplinary bibliographic databases.
• The search strategy has been adapted to each database and platform.
• Appropriate controlled vocabulary has been used where available.
• Free-text terms have been included where appropriate.
• Search dates have been recorded.
• Any language, publication-date, study-design, or other limits have been documented.
• Supplementary searching is described separately from the principal database search where appropriate.
• The complete reproducible search strategies are retained for reporting or supplementary material.

Limitations and important qualifications

The classification of electronic information resources is not always perfectly simple.

Some modern resources combine several functions. PubMed, for example, provides a search interface while simultaneously giving access to a broader collection of NLM bibliographic records that extends beyond MEDLINE.[6] Scopus combines a large citation database with its own integrated search interface. The Cochrane Library contains several types of evidence resources rather than functioning merely as a simple interface.

Database products may also change over time, and institutional subscriptions can differ. The precise database version, available coverage, interface functionality, and indexing status should therefore be verified at the time a systematic-review search is conducted.

Finally, identifying the correct database and platform does not by itself guarantee a comprehensive search. Search quality also depends on the formulation of the research question, selection of appropriate information sources, construction and validation of search concepts, appropriate use of controlled vocabulary and free-text terms, supplementary searching, and transparent reporting.

Key takeaways

1. A database is the collection of information being searched; a platform is the interface used to search it.

2. MEDLINE via Ovid means that MEDLINE is the database and Ovid is the platform.

3. CINAHL via EBSCOhost means that CINAHL is the database and EBSCOhost is the platform.

4. CENTRAL via the Cochrane Library identifies CENTRAL as the database and the Cochrane Library as the environment through which it is searched.

5. PubMed and MEDLINE are not synonymous. MEDLINE is the largest component of PubMed, while PubMed also contains additional records.

6. Embase via Ovid and Embase.com are not two independent databases. They represent access to Embase through different search environments.

7. Google Scholar is a scholarly search engine, whereas ScienceDirect is primarily a publisher-based full-text platform.

8. Different platforms may use different search syntax, while different databases may use different controlled vocabularies.

9. A search strategy should therefore be translated and adapted, rather than simply copied, between databases and platforms.

10. For reproducibility, systematic-review authors should clearly report the database or resource, platform where applicable, search strategy, search date, and any limits or supplementary methods.

References

1. Rethlefsen ML, Kirtley S, Waffenschmidt S, Ayala AP, Moher D, Page MJ, et al. PRISMA-S: an extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Syst Rev. 2021;10(1):39. doi:10.1186/s13643-020-01542-z.

2. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71.

3. National Library of Medicine. MEDLINE: overview. U.S. National Library of Medicine. Available from: https://www.nlm.nih.gov/medline/medline_home.html

4. Elsevier. Embase: the biomedical research database. Available from: https://www.elsevier.com/products/embase

5. EBSCO. CINAHL Database. Available from: https://about.ebsco.com/products/research-databases/cinahl-database

6. National Library of Medicine. About PubMed. Available from: https://pubmed.ncbi.nlm.nih.gov/about/

7. Cochrane. Cochrane Central Register of Controlled Trials (CENTRAL). Cochrane Library. Available from: https://www.cochranelibrary.com/central/about-central

8. Elsevier. Scopus. Available from: https://www.elsevier.com/products/scopus

9. Elsevier. ScienceDirect. Available from: https://www.sciencedirect.com/