Skip to content

Large Language Model-Assisted Clinicians versus Unassisted Clinicians in Clinical Decision Making: A Multi-centre Randomized Controlled Trial in Nairobi, Kenya

Large Language Model-Assisted Clinicians versus Unassisted Clinicians in Clinical Decision Making: A Multi-centre Randomized Controlled Trial in Nairobi, Kenya

Status
Active, not recruiting
Phases
Unknown
Study type
Interventional
Source
PACTR
Registry ID
PACTR202502499779176
Enrollment
9000
Registered
2025-02-24
Start date
2025-03-01
Completion date
Unknown
Last updated
2026-09-14

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Impact of electronic medical record embedded LLM

Interventions

Usual Care i.e. Clinicians without LLM Clinical Decision Support System
LLM Clinical Decision Support System

Sponsors

PATH
Lead Sponsor

Eligibility

Sex/Gender
All

Inclusion criteria

Inclusion criteria: Inclusion Criteria for CLINICIANS: ? Clinicians qualified and registered with the Clinical Officers Council in Kenya ? Clinicians with 3 months prior practice at Penda facility at time of enrollment. ? Willing and able to use the Penda electronic health record for all consultations during study period. ? Willing to consent for their participation in the study and is able to undergo training to adhere to study protocol during consultations. Inclusion Criteria for PATIENT PARTICIPANTS: ? Willing and able to adhere to study requirements, including attendance of follow up consultations. ? Willing to consent for their participation in the study.

Exclusion criteria

Exclusion criteria: Exclusion Criteria for CLINICIANS: ? None. Exclusion Criteria for PATIENT PARTICIPANTS ? Adult patients attending for non-acute/planned wellness visits (for example, weight checks, vaccinations, routine antenatal care). *Paediatric patients attending for any reason are eligible for participation. ? Unable or unwilling to be contacted by phone. ? Adults not of sound mind ? Minors aged 12 years and above who have declined assent and whose caregivers have declined consent for participation in the study. ? Unable or unwilling to be contacted by study staff to collect follow-up outcome data. ? Patients known to be requiring immediate or emergency care.

Design outcomes

Primary

MeasureTime frame
Composite measure of treatment failure, indicated by the occurrence of any of the following within 14 days: I.Re-presentation to Primary Care: Any return visit to a primary care provider (e.g., Penda facility, over-the-counter pharmacy, another primary care facility) with ongoing or unresolved symptoms reported during the first visit. II.Escalation of Care: Any unplanned visit to a secondary care or emergency care provider for an issue related to the initial (primary care) presenting complaint. III.Safety/Adverse Events: Any safety event linked to the initial presentation

Secondary

MeasureTime frame
Safety and appropriateness of LLM-assisted consultations;Effectiveness of the LLM CDSS in the diagnosis and treatment of sentinel conditions (hypertension, type 2 diabetes mellitus, and malnutrition in children under 5 years).;Effectiveness of the LLM CDSS on reducing inappropriate antibiotic and antimalarial prescriptions;Patient satisfaction with consultations based on perceptions of consultation thoroughness, clarity, and overall satisfaction.

Countries

Kenya

Contacts

Public ContactBilal;Bilal Mateen;Mateen

PATH;PATH

bmateena@path.org;bmateena@path.org+12062853500;+12062853500

Outcome results

None listed

Source: PACTR (via WHO ICTRP) · Data processed: Sep 19, 2026