DATA POINT - A HICS INITIATIVE

 


Patterns of Data Dispersion in Medical Statistics


1. Why Dispersion Matters in Medicine

Dispersion describes how spread out data values are around a central point. In clinical research, dispersion helps identify:

  • Biological variability
  • Measurement precision
  • Treatment response differences
  • Subgroups within a population
  • Outliers that may signal pathology or error

2. Symmetry and Skewness in Dispersion

A. Symmetric Dispersion

  • Data spread is equal on both sides of the mean
  • Mean ≈ Median ≈ Mode
  • Often seen in controlled physiological variables
  • Example: Adult male height

B. Asymmetric (Skewed) Dispersion

Right (Positive) Skew

  • Long tail to the right
  • Mean > Median
  • Examples: Length of hospital stay, triglycerides

Left (Negative) Skew

  • Long tail to the left
  • Mean < Median
  • Examples: Age at onset of certain genetic disorders

 


3. Narrow vs. Wide Dispersion

A. Narrow Dispersion

  • Tight clustering of values
  • Indicates homogeneity or strong physiological regulation
  • Example: Blood pH in healthy adults

B. Wide Dispersion

  • Broad spread of values
  • Suggests heterogeneity, disease influence, or measurement variability
  • Example: Blood glucose in diabetic patients

 


4. Uniform, Clustered, and Multimodal Patterns

A. Uniform Dispersion

  • Values evenly distributed
  • Rare in clinical datasets
  • Seen in simulations or randomization processes

B. Clustered Dispersion

  • Data form distinct subgroups
  • Indicates underlying phenotypes or populations
  • Examples:
    • Bimodal trauma patient ages
    • Cholesterol levels in treated vs. untreated groups

C. Modal Patterns

Unimodal: One peak — most biological variables
Bimodal: Two peaks — suggests two populations
Multimodal: Multiple peaks — mixed datasets or phenotypes


5. Outlier‑Driven Dispersion

Sources of Outliers

  • True clinical extremes (e.g., very high CRP in sepsis)
  • Measurement or data entry errors
  • Rare conditions

Impact on Analysis

  • Inflates variance and SD
  • Distorts the mean
  • May require robust statistics (median, IQR) or transformations

 


6. Key Measures of Dispersion

Measure

Best Used For

Notes

Range

Quick sense of spread

Very sensitive to outliers

Interquartile Range (IQR)

Skewed data

Robust; used in boxplots

Variance

Parametric tests

Harder to interpret clinically

Standard Deviation (SD)

Normal distributions

Most common measure

Coefficient of Variation (CV)

Comparing variability across units

Useful in lab medicine


7. Clinical Interpretation of Dispersion

  • High BP variability → increased cardiovascular risk
  • Low assay variability → high precision
  • Wide treatment response variability → need for personalized therapy
  • Clustered patterns → disease subtypes or phenotypes


Comments

Popular posts from this blog

Physiology Note - Respiratory Mechanics during Positive Pressure Ventilation

Physiology Note 1: Perfusion Pressures

Physiology note 2 : Cerebral Autoregulation