This repository packages a portable Ollama/OpenWebUI assistant configuration for clinical research appraisal.
What this is
This is not a fine-tuned model. It is a reusable model wrapper/persona containing:
an Ollama Modelfile
the original OpenWebUI JSON export
a system prompt for structured clinical research appraisal
low-temperature inference settings
Base model
gemma4:latest
Before creating the model locally, make sure the base model is available in Ollama:
ollama pull gemma4:latest
If gemma4:latest is not available on your machine or Ollama registry, edit the FROM line in Modelfile, for example:
FROM gemma3:4b
or another model you have available.
Install locally in Ollama
Clone or download this repository, then run:
bash
1ollama create clinical-research-appraiser -f Modelfile
2ollama run clinical-research-appraiser
Test prompt
Appraise this study abstract. Identify the research question, PICO, design, bias/confounding, statistical concerns, clinical relevance, and suggested improvements.
Use in OpenWebUI
After creating the model in Ollama, open OpenWebUI and select:
clinical-research-appraiser
Alternatively, import the included OpenWebUI JSON if your OpenWebUI version supports importing model/preset exports.
Safety / scope
This assistant is for research appraisal and education. It should not be used as a substitute for clinical judgment, peer review, ethics review, or statistical consultation.
Original system behavior
The assistant is configured to:
Identify the main research question.
Define the PICO/PECO.
Identify exposure/intervention, comparator, and outcome variables.
Classify the study design.
Assess internal validity, external validity, bias, confounding, and missing controls.
Comment on sample size and statistical appropriateness.
