Research hypothesis validation and feasibility assessment skill
Pure instruction skill. LLM medical knowledge + compare_with_literature, list_saved_references, search_local_references, validate_concept。
Triggers: hypothesis, hypothesis, PICO, feasibility, feasibility, testable, research plan assessment
Framework 1: Hypothesis Validation
Step 1: PICO analysis
| Component |
Detection focus |
| Population |
patients, subjects, ICU, surgical... |
| Intervention |
treatment, therapy, drug, procedure... |
| Comparison |
versus, vs, placebo, standard, control... |
| Outcome |
mortality, survival, recovery, score, rate... |
Step 2: Testability
- Directional — higher/lower/better/worse/reduce? ✅ vs descriptive only ❌
- Comparison — significantly/compared/than? ✅ vs missing baseline ❌
- Type — Null (no difference) vs Alternative (directional)
- Specificity — <10 words too vague, 10-50 appropriate, >50 overly complex
Step 3: Study Type Recommendation
| Hypothesis characteristics |
Recommended design |
| effect/efficacy/treatment |
RCT or Cohort |
| risk factor/association |
Cohort or Case-Control |
| prevalence/frequency |
Cross-Sectional |
Framework 2: Feasibility Assessment
Sample Size Quick Check
| Type |
Notes |
| N ≤ 10 |
Case report/series |
| N < 30 RCT |
Pilot or crossover |
| N ≥ 100 |
Mostly reasonable design |
Timeline by Study Type
| Type |
IRB |
Main work |
Analysis and writing |
| RCT |
2 mo |
Recruitment + enrollment ÷3 |
÷4 |
| Retrospective |
1 mo |
Data extraction 2 mo |
÷2 |
| SR |
1 mo (protocol) |
Search + screening ÷3 |
÷3 |
Feasibility Checklist (17 items, 5 categories)
| Category |
Critical Items |
High Items |
| Data & Sample |
Data accessibility, sample reachability |
Inclusion/exclusion criteria |
| Ethics |
IRB plan、informed consent |
privacy |
| Resources |
— |
funding, timeline, team, equipment |
| Methodology |
primary outcome |
stat plan, validated tools, confounders |
| Publication |
— |
target journal, reporting guideline, novelty |
All Critical items must be ✅ before proceeding.
Usage principles
- PICO analysis uses semantic understanding, not keyword matching
- Feasibility assessment should err on the conservative side
- For literature comparison use
compare_with_literature MCP tool
