Tip #70: Building a (Database) Creature with the Best Features: An Unscientific Report

It’s alive! Or we wish it were.

On June 16th, 2026, the Medical Library Association User Experience Caucus hosted the networking event, “Building a (Database) Creature with the Best Features”, at the Library Evidence Synthesis Services Symposium (LESSS). During this event, more than sixty evidence synthesis librarians across disciplines assembled in an unscientific effort to cobble together their dream database platform based on the ones they search every day.

The protocol was prospective, but not IRB approved (come to think of it, what mad scientist has ever bothered with submitting an IRB application?).

A quick overview of our unscientific methods

The event was 90 minutes long, and took place over Zoom. Data were collected using  "The secret laboratory of Dr. Frankenstein, MLS" Padlet

Data were collected for the following categories:

• Exporting references
• Display settings
• Field searching
• Search history
• Wildcards / truncation
• Proximity searching
• Thesauri
• Missing body parts (any features we missed)
• Best database overall
•  Illegal bonus round: Platform most likely to be chased with flaming torches

For each category, participants chose the platform that had their favorite feature, and commented what that favorite feature was under the platform. They also had the option to unmute and “pitch” their favorite feature

Participants chose from 16 preselected platforms:

1. ACM Digital Library
2. Cochrane Library (Wiley)
3. EBSCOHost
4. Embase.com
5. First Search
6. Google Scholar
7. IEEE Xplore
8. JSTOR
9. The Lens
10. OpenAlex
11. Ovid
12. ProQuest
13. PubMed
14. Scopus
15. Web of Science
16. WHO Global Index Medicus

At the end of each category discussion, participants “voted” for their favorite platform(s) for the category using the thumbs up emoji. Note that participants were allowed to vote for multiple platforms within a single category.

For more details into our methods, check out our unscientific protocol.

The results are in. Pursue them if you dare! Muahahahaha!

An overview of our mad scientists

129 aspiring mad scientists registered for the event, each representing a wide array of disciplines. Most registrants represented multiple disciplines, and a few even stated they served “all” disciplines. It turns out librarians tend to wear many hats. Who knew? (Yes, we ask this with the utmost sarcasm.)

Unsurprisingly, the health sciences was the most represented discipline followed closely by the social sciences; however, evil geniuses convened from a wealth of other disciplines, as well, as shown in the word cloud below. An accessible version of the word cloud in table format can be found in this Google Sheet. Note that some disciplines may overlap (mad scientists aren’t renowned for being systematic).

Of the 129 aspiring mad scientists that registered for the event, just over 60 were brave enough (or foolish enough?) to take up the scalpel.

An Overview of the Results and Our Resulting Blueprint

An overview of the voting results by category is provided below in the forms of a bar chart and a table. These results can also be accessed via this Google Sheet. Note that scientists were allowed to comment and vote on multiple platforms within a single category. As alluded to in the overview of our mad scientists, most of our scientists derived from health sciences fields (to be fair, most mad scientists derive from the health sciences. Don’t fact check us on that statistic. Just trust us on this one 😉); in consequence, the results are going to be biased towards that field.

Figure 1: Bar chart of voting results for 16 platforms by category

Table1: Table of voting results for 16 platforms by category

Database	A. Exporting References	B. Display Settings	C. Field Searching	D. Search History	E. Wildcards and Truncation	F. Proximity Searching	G. Thesauri	H. Bonus: Missing Body Parts
ACM Digital Library	0	0	0	0	0	0	0	0
Cochrane Library (Wiley)	4	6	1	6	1	4	0	1
EBSCOhost	6	5	9	0	0	10	8	2
Embase.com	14	12	9	12	4	7	17	2
FirstSearch	1	0	0	0	0	0	0	0
Google Scholar	1	4	3	0	0	0	0	0
IEEE Xplore	0	0	0	0	0	0	0	0
JSTOR	0	0	0	0	0	0	0	0
The Lens	0	0	0	0	1	0	0	0
OpenAlex	3	1	0	0	0	0	0	1
Ovid	19	21	22	22	15	16	12	8
ProQuest	1	2	2	0	0	0	1	0
PubMed	23	20	11	16	6	1	24	13
Scopus	11	6	5	1	3	6	0	1
Web of Science	7	11	14	7	12	5	0	3
WHO Global Index Medicus	1	0	0	0	0	0	0	0

Below is our resulting blueprint for our database creature with the best features. To accommodate our more squeamish readers (and save on cleaning supplies), instead of stitching together body parts we opted to stitch together a swanky outfit for our creature.

Ovid and PubMed were our winners across platform features, with Ovid being the winner for 5 out of the 8 (62.5%) categories and PubMed for 3 out of the 8 (37.5%) categories. Consequently, we only had 2 colors to work with (blue for Ovid, and yellow for PubMed). Even with this limitation, our monster looks pretty snazzy, if we do say so ourselves! And who are we to argue with our monster’s winning smile? (Seriously, don’t argue with him. As all librarians know, database creatures are not the most well-mannered of beasts. And, we, like any evil scientist, will not take responsibility for our creature’s actions.)

Figure 2: Database creature with the best features

 

Dissecting the Categories

Get out your scalpels and forceps! In this section we provide a deeper look at the results for each of the categories. Once more, note that scientists were allowed to comment and vote for multiple platforms within a single category. Additionally, due to our composition of mad scientists, the results will be biased towards the health sciences.

For the full list of comments for each category, check out the Padlet!

A. Exporting References

Mad scientists were asked which platform had the best features for exporting references. Of the 16 platforms, PubMed was the winner for that category with 23 votes. Ovid was in second place with 19 votes, and Embase.com was in third place with 14.

With regards to exporting references, some of the winning features for PubMed included:

• The ability to export up to 10,000 references at once
• Not needing to sign in to export results
• The ability to directly export results (rather than receiving results via email)
• The ability to export PMIDs
• The ability to get data via the Entrez API

B. Display Settings

Mad scientists were asked which platform had the best display setting features. Ovid came in first with 21 votes, followed by PubMed with 20 votes, and Embase.com with 12 votes.

With regards to display settings, some of the winning features of Ovid included:

• The prominent display of one’s search history
• Easy access to search editing and linking functions
• The ability to toggle the abstract view for individual results

C. Field Searching

Mad scientists were asked which platform had the best features for field searching. Ovid took first place at 22 votes, Web of Science took second at 14 votes, and PubMed took third at 11 votes.

With regards to field searching, some of Ovid’s winning features included:

• The ability to use several fields at once
• The ability to add or remove field codes behind an entire expression
• The provision of a list of available field codes
• Ease of searching indexes
• The wide array and granularity of field options
• The ability to work with Polyglot
• The helpfulness of the search fields form and database field guide
• The manual distinguishes between word and phrase indexed fields
• The ability to search the keyword heading field separately from the subject heading field

D. Search History

Mad scientists were asked which platform had the best search history features. Ovid came in first with 22 votes, followed by PubMed with 16, and Embase.com with 12.

With regards to search history, some of Ovid’s winning features included:

• The search history’s high visibility
• Easy search export and edit functions
• The ability to save searches
• Only needing one click to copy search details
• The ability to link to search history without needing to sign in
• The potential for search history to serve as a pedagogical tool for new searchers
• Automatic renumbering of searches in the search history when lines are deleted
• The Ovid History Jumpstart tool
• Easy line combinations
• Easy conversion of search history copied from Ovid into tables in Word
• The structure facilitates search term testing
• The ability to copy detailed search histories into plain text

E. Wildcards and Truncation

Mad scientists were asked which platform had the best wildcard and truncation features. Ovid had the most votes with 15, followed by Web of Science with 12, and PubMed with 6.

With regards to wildcards and truncation, some of Ovid’s winning features included:

• The availability of both limited and unlimited truncation
• The ability to use truncation within phrases
• The flexibility of truncation and wildcards
• The ability to use truncation to hack stopwords (e.g., searching “the? public” retrieves things like “the public”)
• The ability to specify matching of all (*), 1 or 0 (?), and exactly 1 character (#)

F. Proximity Searching

Mad scientists were asked which platform had the best proximity searching features. Yet again, Ovid came in first with 16 votes, followed by EBSCOhost with 10, and Embase.com with 7.

With regards to proximity searching, some of Ovid’s winning features included:

• Easy of usability and edit functions
• The ability to bundle together groups of words to search in proximity with one another

Though we instructed our scientists to keep things positive, we couldn’t blame one of our scientists for lamenting how proximity interpretations can differ between databases (e.g., ADJ 3 was noted as being the same as N2 in EBSCOhost), nor another for the snide remark that all databases’ proximity searching features were at least superior to PubMed’s (though one optimistic scientist praised PubMed for at least having a proximity feature now).

G. Thesaurus

Mad scientists were asked which platform had the best thesaurus features. PubMed came in first this time with 24 votes, followed by Embase.com with 17, and Ovid with 12.

With regards to thesauri, some of PubMed’s winning features included:

• The intuitiveness of the MeSH database
• The MeSH tree view
• Transparency in MeSH terms (including definitions, scope notes, etc.)
• The presence of an external subject heading site
• Being able to access the MeSH database without needing to sign in
• The presence and usability of subheadings
• The ability to search subheadings without specifying a main subject heading
• The presence of entry terms

 One scientist also gave a plug for PubMed’s phrase index.

H. Bonus Round: Missing Body Parts

In this round we asked mad scientists if there were any favorite platform features we missed. In this vote, PubMed came up on top with 13 votes, followed by Ovid with 8, and Web of Science with 3.

For PubMed, some of these favorite features included:

• The similar articles feature
• That it’s free to search
• It’s an internationally recognized database
• The cited by feature
• The search details, which shows you the location of your search errors
• The existence of subheadings and floating subheadings
• Facility of combining searches in the search history
• The ability to enter a  string of PMIDs without syntax (i.e., ORs) to retrieve the corresponding records
• The ability to display the results as a list of PMIDs
• The availability of the API (Entrez Direct / E-Utilities)
• The clipboard feature
• Automatic term mapping serving as a great (though underestimated) source of search terms

Final Vote: Best and Worst Platforms

As a fitting end to our experiment, we also had our mad scientists vote for their overall favorite platform…and also the platform they would most like to chase down with burning torches (we didn’t exclude evil scientists from our experiment, after all).

First, the results for the best platform. As shown in the bar chart below, Ovid unsurprisingly came in first with 17 votes, with PubMed coming in second at 13, and Web of Science in third at 7.

Now the worst platform. Following a rather cathartic venting session, the razzie ultimately went to EBSCOhost with 11 votes. Google Scholar was awarded second place with 8 votes, and Proquest and Scopus tied for third place at 7 votes.

Finis

What are your favorite platform features, or what platform would you chase with a burning torch? (Either one is fine; we have no moral compass here). Let us know in the comments!

A special thanks to the LESSS coordinating team for hosting our session, all the mad scientists who participated, and our colleagues who acted as guinea pigs when we first piloted our experiment!

Note: The fabulous Frankenstein’s monster image was modified from Canva.