zhongwei/gh-k-dense-ai-claude-scientific-writer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit

Browse Source
This commit is contained in:
Zhongwei Li
2025-11-30 08:30:14 +08:00
commit 1dd5bee3b4
335 changed files with 147360 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

530
skills/clinical-reports/references/data_presentation.md Normal file
View File

@@ -0,0 +1,530 @@
# Data Presentation in Clinical Reports
## Tables for Clinical Data
### Table Design Principles
**General guidelines:**
- Clear, concise title describing table contents
- Column headers with units
- Row labels aligned left, data aligned appropriately (numbers right, text left)
- Footnotes for abbreviations, statistical notation, special cases
- Consistent decimal places (typically 1-2 for percentages, 1-3 for continuous variables)
- Consistent formatting throughout document
**Title placement:**
- Above table
- Numbered sequentially (Table 1, Table 2, etc.)
- Descriptive enough to stand alone
**Footnote symbols (in order):**
- *, †, ‡, §, ||, ¶, #
- Or use superscript letters (a, b, c...)
- Or use superscript numbers if not confused with references
### Demographic and Baseline Characteristics Table
**Purpose:** Describe study population at baseline
**Standard format:**
```
Table 1. Baseline Demographics and Clinical Characteristics
Characteristic Treatment Group Control Group Total
(N=150) (N=145) (N=295)
─────────────────────────────────────────────────────────────────────────
Age, years
Mean (SD) 64.2 (8.5) 63.8 (9.1) 64.0 (8.8)
Median (IQR) 65 (58-71) 64 (57-70) 64 (58-71)
Range 45-82 43-85 43-85
Sex, n (%)
Male 95 (63.3) 88 (60.7) 183 (62.0)
Female 55 (36.7) 57 (39.3) 112 (38.0)
Race, n (%)
White 110 (73.3) 105 (72.4) 215 (72.9)
Black/African American 25 (16.7) 28 (19.3) 53 (18.0)
Asian 10 (6.7) 8 (5.5) 18 (6.1)
Other 5 (3.3) 4 (2.8) 9 (3.0)
BMI, kg/m²
Mean (SD) 28.5 (4.2) 28.1 (4.5) 28.3 (4.4)
Baseline HbA1c, %
Mean (SD) 8.9 (1.2) 9.0 (1.3) 9.0 (1.2)
Disease duration, years
Median (IQR) 6 (3-10) 5 (3-9) 6 (3-10)
Prior medications, n (%)
Metformin 135 (90.0) 130 (89.7) 265 (89.8)
Sulfonylurea 45 (30.0) 42 (29.0) 87 (29.5)
Insulin 20 (13.3) 18 (12.4) 38 (12.9)
─────────────────────────────────────────────────────────────────────────
SD = standard deviation; IQR = interquartile range; BMI = body mass index;
HbA1c = hemoglobin A1c
```
**Key elements:**
- Sample size for each group (N=)
- Continuous variables: mean (SD), median (IQR), range
- Categorical variables: n (%)
- No p-values for baseline comparisons (debated but generally not recommended)
### Efficacy Results Table
**Purpose:** Present primary and secondary endpoint results
**Example:**
```
Table 2. Primary and Secondary Efficacy Endpoints at Week 24
Endpoint Treatment Control Difference P-value
(N=150) (N=145) (95% CI)
──────────────────────────────────────────────────────────────────────────────────
Primary Endpoint
Change in HbA1c from baseline, %
Mean (SE) -1.8 (0.1) -0.6 (0.1) -1.2 <0.001
95% CI (-2.0, -1.6) (-0.8, -0.4) (-1.5, -0.9)
Secondary Endpoints
Change in FPG, mg/dL
Mean (SE) -42.5 (3.2) -15.2 (3.4) -27.3 <0.001
95% CI (-48.8, -36.2) (-21.9, -8.5) (-36.4, -18.2)
% achieving HbA1c <7%
n (%) 78 (52.0) 25 (17.2) - <0.001
95% CI (43.9, 60.1) (11.4, 24.5)
Change in body weight, kg
Mean (SE) -3.2 (0.4) -0.5 (0.4) -2.7 <0.001
95% CI (-4.0, -2.4) (-1.3, 0.3) (-3.8, -1.6)
──────────────────────────────────────────────────────────────────────────────
SE = standard error; CI = confidence interval; HbA1c = hemoglobin A1c;
FPG = fasting plasma glucose
```
**Statistical presentation:**
- Point estimates with measures of precision (SE or CI)
- p-values (consider adjustment for multiplicity)
- Effect size (difference or ratio) with 95% CI
- Significance level noted (e.g., p<0.05, p<0.01, p<0.001)
### Adverse Events Table
**Purpose:** Summarize safety data
**Example:**
```
Table 3. Summary of Adverse Events
Event Category Treatment Control P-value
(N=150) (N=145)
n (%) n (%)
──────────────────────────────────────────────────────────────────────────
Any adverse event 120 (80.0) 95 (65.5) 0.004
Treatment-related adverse events 85 (56.7) 42 (29.0) <0.001
Serious adverse events 12 (8.0) 8 (5.5) 0.412
Adverse events leading to 8 (5.3) 4 (2.8) 0.257
discontinuation
Deaths 0 (0.0) 1 (0.7) 0.492
Common adverse events (≥5% in any group)
Nausea 45 (30.0) 12 (8.3) <0.001
Diarrhea 38 (25.3) 10 (6.9) <0.001
Headache 22 (14.7) 18 (12.4) 0.568
Hypoglycemia 18 (12.0) 5 (3.4) 0.007
Dizziness 12 (8.0) 8 (5.5) 0.412
──────────────────────────────────────────────────────────────────────────
Adverse events coded using MedDRA version 24.0
```
**Key elements:**
- Overall AE summary
- Serious AEs highlighted
- Deaths reported
- Common AEs (typically ≥5% or ≥10% threshold)
- MedDRA coding indicated
### Laboratory Abnormalities Table
**Shift tables showing changes from baseline:**
```
Table 4. Laboratory Values Meeting Predefined Criteria for Abnormality
Laboratory Parameter Treatment Control
(N=150) (N=145)
n (%) n (%)
──────────────────────────────────────────────────────────────────────────
ALT >3× ULN 8 (5.3) 3 (2.1)
AST >3× ULN 5 (3.3) 2 (1.4)
Total bilirubin >2× ULN 2 (1.3) 1 (0.7)
Creatinine >1.5× baseline 12 (8.0) 5 (3.4)
Hemoglobin <10 g/dL 3 (2.0) 2 (1.4)
Platelets <100 × 10³/μL 1 (0.7) 0 (0.0)
──────────────────────────────────────────────────────────────────────────
ULN = upper limit of normal; ALT = alanine aminotransferase;
AST = aspartate aminotransferase
```
### Patient Disposition Table (CONSORT Format)
```
Table 5. Patient Disposition
Disposition Treatment Control Total
(N=150) (N=145) (N=295)
────────────────────────────────────────────────────────────────────────────
Screened - - 425
Randomized 150 145 295
Completed study 135 (90.0) 130 (89.7) 265 (89.8)
Discontinued, n (%) 15 (10.0) 15 (10.3) 30 (10.2)
Adverse event 8 (5.3) 4 (2.8) 12 (4.1)
Lack of efficacy 2 (1.3) 5 (3.4) 7 (2.4)
Lost to follow-up 3 (2.0) 4 (2.8) 7 (2.4)
Withdrawal of consent 2 (1.3) 2 (1.4) 4 (1.4)
Included in efficacy analysis
ITT population 150 (100) 145 (100) 295 (100)
Per-protocol population 142 (94.7) 138 (95.2) 280 (94.9)
Included in safety analysis 150 (100) 145 (100) 295 (100)
────────────────────────────────────────────────────────────────────────────
ITT = intent-to-treat
```
## Figures for Clinical Data
### Figure Design Principles
**General guidelines:**
- Clear, concise caption/legend below figure
- Numbered sequentially (Figure 1, Figure 2, etc.)
- Axis labels with units
- Legible font size (minimum 8-10 point)
- High resolution (300 dpi for print, 150 dpi for web)
- Color-blind friendly palette
- Black and white compatible (use different symbols/patterns)
**Figure caption:**
- Describes what is shown
- Explains symbols, error bars, statistical annotations
- Defines abbreviations
- Provides context for interpretation
### CONSORT Flow Diagram
**Purpose:** Show patient flow through randomized trial
```
Assessed for eligibility (n=425)
│
┌─────────────────────┴─────────────────────┐
│ │
Excluded (n=130) │
• Not meeting inclusion criteria (n=85) │
• Declined to participate (n=32) │
• Other reasons (n=13) │
│
Randomized (n=295)
│
┌───────────────────────────────┴───────────────────────────────┐
│ │
Allocated to Treatment (n=150) Allocated to Control (n=145)
• Received allocated intervention (n=148) • Received allocated intervention (n=143)
• Did not receive allocated intervention (n=2) • Did not receive allocated intervention (n=2)
Reasons: withdrew consent before treatment Reasons: withdrew consent before treatment
│ │
┌───────────┴────────────┐ ┌──────────────┴─────────────┐
│ │ │ │
Lost to follow-up (n=3) Discontinued (n=12) Lost to follow-up (n=4) Discontinued (n=11)
• Adverse events (n=8) • Adverse events (n=4)
• Lack of efficacy (n=2) • Lack of efficacy (n=5)
• Withdrew consent (n=2) • Withdrew consent (n=2)
│ │
Analyzed (n=150) Analyzed (n=145)
• ITT analysis (n=150) • ITT analysis (n=145)
• Per-protocol analysis (n=142) • Per-protocol analysis (n=138)
• Excluded from analysis (n=0) • Excluded from analysis (n=0)
```
### Kaplan-Meier Survival Curve
**Purpose:** Show time-to-event data
**Elements:**
- X-axis: Time (weeks, months, years)
- Y-axis: Probability of event-free survival (0 to 1 or 0% to 100%)
- Separate curves for each treatment group
- Censored observations marked (often with vertical tick marks)
- Number at risk table below graph
- Median survival time indicated
- Log-rank p-value
- Hazard ratio with 95% CI
**Caption example:**
```
Figure 1. Kaplan-Meier Curves for Overall Survival
Kaplan-Meier estimates of overall survival in the treatment and control groups.
Tick marks indicate censored observations. Number at risk shown below graph.
Log-rank p<0.001. Median survival: Treatment 24.5 months (95% CI: 22.1-26.8),
Control 18.2 months (95% CI: 16.5-20.1). Hazard ratio 0.68 (95% CI: 0.55-0.84).
```
### Forest Plot
**Purpose:** Display subgroup analyses or meta-analysis results
**Elements:**
- Point estimates (squares or diamonds)
- Size of symbol proportional to precision (inverse variance) or sample size
- Horizontal lines showing 95% CI
- Vertical line at null effect (HR=1.0, OR=1.0, or difference=0)
- Subgroup labels on left
- Effect size values on right
- Overall estimate (if meta-analysis)
- Heterogeneity statistics (I², p-value)
**Caption example:**
```
Figure 2. Forest Plot of Treatment Effect by Subgroup
Effect of treatment vs. control on primary endpoint across pre-specified subgroups.
Squares represent point estimates; horizontal lines represent 95% confidence intervals.
Square size is proportional to subgroup sample size. Overall effect shown as diamond.
p-value for interaction testing heterogeneity of treatment effect across subgroups.
```
### Box Plot
**Purpose:** Show distribution of continuous variable
**Elements:**
- Box: IQR (25th to 75th percentile)
- Line in box: Median
- Whiskers: Extend to most extreme data point within 1.5 × IQR
- Outliers: Points beyond whiskers (often shown as circles)
- X-axis: Groups or time points
- Y-axis: Continuous variable with units
### Scatter Plot with Regression
**Purpose:** Show relationship between two continuous variables
**Elements:**
- X-axis: Independent variable
- Y-axis: Dependent variable
- Individual data points
- Regression line (if appropriate)
- Regression equation
- R² value
- P-value for slope
- 95% confidence interval for regression line (optional, shown as shaded area)
### Spaghetti Plot
**Purpose:** Show individual trajectories over time
**Elements:**
- X-axis: Time
- Y-axis: Outcome variable
- Individual patient lines (often semi-transparent)
- Mean trajectory (bold line)
- Separate colors for treatment groups
### Bar Chart
**Purpose:** Compare proportions or means across groups
**Elements:**
- Clear separation between bars
- Error bars (SEM or 95% CI)
- Y-axis starts at 0 (do not truncate for bar charts)
- Group labels on X-axis
- Value labels on Y-axis with units
- Statistical significance indicated (p-values or asterisks)
**Avoid:**
- 3D bar charts (distort perception)
- Excessive decoration
- Truncated Y-axis for bars
### Line Graph
**Purpose:** Show changes over time
**Elements:**
- X-axis: Time (with consistent intervals)
- Y-axis: Outcome variable
- Separate lines for each group (different colors/patterns)
- Data points marked (circles, squares, triangles)
- Error bars at each time point (SE or 95% CI)
- Legend identifying groups
- Grid lines (optional, light gray)
### Histogram
**Purpose:** Show distribution of continuous variable
**Elements:**
- X-axis: Variable (divided into bins)
- Y-axis: Frequency or density
- Appropriate bin width (not too few, not too many)
- Overlay normal distribution curve (if testing normality)
## Special Considerations for Clinical Data
### Presenting Proportions
**Numerator and denominator:**
- Always provide both: 25/100 (25%)
- Not just percentage (25%)
**Percentages:**
- No decimal places if n<100
- 1 decimal place if n≥100
- Never report >1 decimal place for percentages
**Confidence intervals for proportions:**
- Wilson score interval or exact binomial (better than Wald for small samples)
- Always report with percentage
### Presenting Continuous Data
**Measures of central tendency:**
- Mean for normally distributed data
- Median for skewed data or ordinal data
- Report both if distribution unclear
**Measures of dispersion:**
- **Standard deviation (SD)**: Describes variability in data
- **Standard error (SE)**: Describes precision of mean estimate
- **95% Confidence interval**: Preferred for inferential statistics
- **Interquartile range (IQR)**: With median for skewed data
- **Range**: Min to max
**When to use each:**
- Descriptive statistics → Mean (SD) or Median (IQR)
- Inferential statistics → Mean (95% CI) or Mean (SE)
- Never use ± without specifying SD, SE, or CI
### Presenting P-values
**Reporting guidelines:**
- Report exact p-values to 2-3 decimal places (p=0.042)
- For very small p-values, use p<0.001 (not p=0.000)
- Do not report as "NS" or "p=NS"
- For non-significant results, report exact p-value (p=0.18, not p>0.05)
- Specify two-tailed unless pre-specified one-tailed
- Correct for multiple comparisons when appropriate
- Report significance threshold used (α=0.05 is standard)
**Avoid:**
- p<0.05 (report exact value)
- p=0.00 (impossible)
- Multiple decimal places (p=0.04235891)
### Statistical Significance Indicators
**Options:**
1. Report p-values in table
2. Use asterisks with legend:
- *p<0.05
- **p<0.01
- ***p<0.001
3. Use confidence intervals (preferred)
### Confidence Intervals
**Reporting:**
- 95% CI is standard
- Format: (lower limit, upper limit)
- Or: lower limit to upper limit
- Or: lower limit-upper limit
**Interpretation:**
- If CI for difference excludes 0 → significant
- If CI for ratio excludes 1 → significant
- Width of CI indicates precision
### Missing Data
**Indicate clearly:**
- Footnote explaining missing data
- State clearly if analysis is complete case
- Describe imputation method if used
- Report amount of missing data per variable
### Decimal Places and Rounding
**General rules:**
- Report to level of measurement precision
- Consistent decimal places within table
- Round p-values to 2-3 decimal places
- Round percentages to 0-1 decimal place
- Round means/medians to 1-2 decimal places
- Include appropriate significant figures
## Software for Creating Figures
**Statistical software:**
- R (ggplot2) - highly customizable
- GraphPad Prism - user-friendly for biomedical
- SAS, Stata, SPSS - comprehensive statistical packages
- Python (matplotlib, seaborn) - flexible and powerful
**General graphics software:**
- Adobe Illustrator - professional publication-quality
- Inkscape - free vector graphics editor
- PowerPoint - basic graphs, easy to use
- BioRender - biological schematics and figures
## Color Schemes
**Color-blind friendly palettes:**
- Avoid red-green combinations
- Use blue-orange, blue-yellow
- Include shape/pattern differences
- Test figures in grayscale
**Recommended palettes:**
- ColorBrewer (designed for data visualization)
- Viridis (perceptually uniform)
- IBM Color Blind Safe Palette
## Image Quality Standards
**Resolution:**
- 300 dpi for print publication
- 150 dpi for web/screen
- Vector graphics (PDF, SVG) preferred for graphs
**File formats:**
- TIFF or EPS for print
- PNG for web
- PDF for vector graphics
- JPEG acceptable for photographs (high quality)
**Image editing:**
- No manipulation that alters data
- Only acceptable adjustments: brightness, contrast, color balance applied to entire image
- Document all adjustments
- Provide original images if requested
---
This reference provides comprehensive guidance for presenting clinical data in tables and figures following best practices and publication standards. Use these guidelines to create clear, accurate, and professional data presentations.