End-to-End Research Workflow in CareFrame

Overview

CareFrame enables a complete research workflow from initial hypothesis formulation to validated evidence sharing. This guide demonstrates how researchers can use the platform to conduct a full research study while maintaining data integrity, methodological rigor, and collaborative potential.

Phase 1: Research Planning

Step 1: Define Research Objectives

1. Navigate to the Planning section
2. Click the Add Objective button
3. Enter your primary research question
4. Select the appropriate objective type (e.g., Research Question)
5. Add any sub-objectives that break down the main question

Example:

• Primary objective: "Investigate the relationship between sleep quality and cardiovascular outcomes"
• Sub-objective 1: "Identify markers of sleep quality most predictive of cardiovascular events"
• Sub-objective 2: "Quantify the relationship between sleep duration and blood pressure"

Step 2: Create Testable Hypotheses

For each objective or sub-objective:

1. Select the objective on the canvas
2. Click Add Hypothesis
3. Write a clear, testable hypothesis statement
4. Define the variables and expected relationship

Example:

• Hypothesis 1: "Lower sleep efficiency (<85%) is associated with higher risk of cardiovascular events in adults over 50"
• Hypothesis 2: "Sleep duration has a U-shaped relationship with systolic blood pressure, with optimal range at 7-8 hours"

Step 3: Generate Alternative Hypotheses

To ensure robust testing:

1. Select an objective
2. Click Auto-Generate Hypotheses
3. Review the AI-generated alternatives
4. Keep the most promising candidates
5. Refine as needed

Phase 2: Literature Evidence

Step 1: Systematic Literature Search

1. Navigate to the Literature Search section
2. Enter search terms related to your hypotheses
3. Apply filters for publication date, study type, etc.
4. Execute search across connected databases
5. Review and save relevant papers

Example:

• Search terms: "sleep quality AND cardiovascular risk OR cardiovascular outcome"
• Filters: Human studies, last 10 years, English language

Step 2: Paper Ranking and Evidence Extraction

1. Navigate to the Ranking tab
2. Sort papers by relevance to your hypotheses
3. Select high-quality papers for detailed review
4. Extract specific claims relevant to your hypotheses
5. Annotate evidence strength and relationship to hypotheses

Example:

• Claim: "In a prospective cohort study of 2,300 adults, sleep efficiency <80% was associated with 27% higher risk of cardiovascular events (HR 1.27, 95% CI 1.12-1.44)"
• Relevance: Directly supports Hypothesis 1
• Strength: Strong (large sample, prospective design)

Phase 3: Data Collection and Processing

Step 1: Connect to Data Sources

1. Navigate to the Data Sources section
2. Select the appropriate data source type
3. Configure connection details
4. Preview and validate the imported data
5. Assign variable types and measurement levels

Example:

• Source: CSV file from sleep study
• Variables: participant_id, age, gender, sleep_efficiency, sleep_duration, systolic_bp, cardiovascular_events

Step 2: Data Cleaning

1. Navigate to the Clean section
2. Identify missing or anomalous values
3. Apply appropriate cleaning methods
4. Document all transformations
5. Validate the cleaned dataset

Example:

• Replace out-of-range values (e.g., negative sleep duration)
• Impute missing blood pressure values using appropriate methods
• Remove duplicate participant records

Step 3: Data Preparation

1. Navigate to the Reshape and Filter sections
2. Transform data into analysis-ready format
3. Create calculated variables if needed
4. Filter to the appropriate analysis sample
5. Document selection criteria

Example:

• Create sleep_efficiency_category variable (<80%, 80-85%, >85%)
• Filter to participants with complete follow-up data
• Calculate follow-up duration variable

Phase 4: Statistical Analysis

Step 1: Model Selection

1. Navigate to the Model section
2. Select the hypothesis to test
3. Choose appropriate statistical test based on:
   • Variable types
   • Distribution characteristics
   • Research question

Example:

• For Hypothesis 1: Cox proportional hazards regression
• For Hypothesis 2: Polynomial regression with quadratic term

Step 2: Assumption Checking

1. Navigate to the Assumptions section
2. Review automatic assumption checks
3. Address any violations through:
   • Data transformation
   • Alternative test selection
   • Robust methods

Example:

• Check proportional hazards assumption for Cox model
• Verify normality of residuals for regression models
• Test for multicollinearity among predictors

Step 3: Running Tests and Interpretation

1. Execute the statistical test
2. Review results output and visualizations
3. Navigate to the Interpret section
4. Document the findings relative to the hypothesis
5. Note limitations and confidence in results

Example:

• "Sleep efficiency <80% was associated with increased cardiovascular risk (HR=1.32, p=0.01), supporting our hypothesis"
• "Sleep duration showed a U-shaped relationship with SBP, with minimum at 7.3 hours (p<0.001)"

Phase 5: Evidence Storage and Validation

Step 1: Submit Results to Blockchain

1. Navigate to the Evidence section
2. Select the hypothesis with new evidence
3. Click Push to Blockchain to submit the hypothesis
4. Click Push Model Evidence to add statistical test results
5. Click Push Literature Evidence to add literature support

Example:

• Submit Hypothesis 1 with metadata
• Add Cox regression results as model evidence
• Add 3 key literature references supporting the finding

Step 2: Validation

1. Team validators navigate to Validator Management
2. Review the pending evidence blocks
3. Validate the evidence if methodologically sound
4. Add any validation notes or concerns
5. Once validation threshold is met, evidence is confirmed

Phase 6: Knowledge Sharing

Step 1: Network Configuration

1. Navigate to the Network section
2. Start server in Host Mode or connect to existing host
3. Configure topics for evidence sharing
4. Establish connections with collaborator instances

Step 2: Evidence Publication

1. Return to the Network section
2. Select the "evidence" topic
3. Reference the validated blockchain evidence
4. Publish to the network
5. Verify receipt by subscribed instances

Phase 7: Protocol Development

Step 1: Protocol Creation Based on Findings

1. Navigate to the Protocol section
2. Create new protocol informed by validated evidence
3. Reference blockchain evidence IDs directly
4. Include literature and statistical evidence
5. Develop comprehensive methodology

Step 2: Protocol Distribution

1. Return to Network section
2. Select the "protocol" topic
3. Share the new protocol with collaborating centers
4. Collect feedback via the same channel

Conclusion

This end-to-end workflow demonstrates how CareFrame integrates traditionally siloed research phases into a seamless, transparent process. By connecting planning, literature, data, analysis, and evidence sharing, CareFrame helps researchers:

1. Maintain consistent focus on research objectives
2. Build upon existing evidence systematically
3. Ensure methodological rigor through appropriate testing
4. Create verifiable, immutable evidence records
5. Share validated knowledge efficiently

Each phase produces outputs that become inputs to subsequent phases, creating a continuous learning system where evidence builds upon evidence in a structured, verifiable manner.