Usability of selected databases for low-resource clinical decision support

Area of Science:

• Medical Informatics
• Global Health
• Clinical Decision Support

Background:

• Smartphones are vital for clinical decision support, but access is limited by cost and coverage in many regions.
• txt2MEDLINE offers medical evidence via text messaging, addressing accessibility challenges.
• Previous evaluations exist, but comparisons with other text-based tools were lacking.

Purpose of the Study:

• To compare the effectiveness of txt2MEDLINE against other text-based databases for clinical queries in resource-limited environments.

Main Methods:

• Searched txt2MEDLINE and five other platforms (askMEDLINE, Cochrane, DynaMed, PubMed PICO, UpToDate) using optimized search strategies.
• Submitted five benchmark queries to each database, truncating results to simulate text message delivery (max 3 messages).
• Clinicians with experience in low-resource settings rated the usefulness of the results on a 5-point Likert scale.

Main Results:

• Median scores (txt2MEDLINE: 3.2, askMEDLINE: 3.2, Cochrane: 3.8, DynaMed: 3.6, PubMed PICO: 3.6, UpToDate: 4.0) indicated no significant one-point differences.
• PubMed PICO (p=0.005) and UpToDate (p=0.002) showed statistically significant improvements over txt2MEDLINE.
• The observed improvements for PubMed PICO (0.4) and UpToDate (0.8) were relatively small on the Likert scale.

Conclusions:

• No single text-based tool demonstrated a one-point advantage in usefulness for clinical decision support.
• txt2MEDLINE is a viable alternative for texting-only clinical decision support in low-resource settings.
• While PubMed PICO and UpToDate offered marginal benefits, txt2MEDLINE's comparability makes it suitable where internet access is restricted.