Making Sense of Medical Data
Descriptive vs. Inferential Statistics and Why They Matter
Picture a study of 200 ICU patients. You’ve got their ages, their diagnoses, whether they survived, how long they stayed. That’s a pile of data - but data on its own doesn’t tell you anything. You have to do something with it, and in medicine, that “something” falls into two buckets: descriptive and inferential statistics.
Descriptive vs. Inferential Statistics: What’s Actually the Difference?
Statistics helps you summarize, analyze, and draw conclusions from data - but not all statistical methods are doing the same job.
Descriptive Statistics – “Telling the Story of the Data”
Summarizes and organizes the raw numbers in front of you.
Gives you an overview without claiming anything beyond that.
Example: “The average age of our 200 ICU patients was 64, and 30% were current smokers.” That’s it - just a snapshot of this group.
Inferential Statistics – “Making Predictions and Testing Hypotheses”
Takes a sample and uses it to make a claim about a larger population.
Lets you test a hypothesis and put a number on how likely something is to be real versus chance.
Example: Using outcomes from those same 200 patients to estimate how a new sedation protocol would perform across ICUs generally, not just for the patients you happened to study.
Key Difference?
Descriptive statistics tells you what’s in front of you. Inferential statistics uses what’s in front of you to make a claim about what you haven’t seen yet.
Types of Data: The Foundation of Statistical Analysis
Before you can run any statistical test, you need to know what kind of data you’re actually holding, because the type of data determines which tests are even valid to use.
Categorical (Qualitative) Data: data that sorts patients into groups rather than measuring them on a scale.
Subtypes:
Nominal Data: categories with no natural order - blood type, eye color, gender.
Ordinal Data: categories that do have an order - cancer stages I through IV, for instance.
Back to our ICU patients: disease status (positive/negative) and medication type (antibiotic vs. antiviral) are both categorical.
Numerical (Quantitative) Data: data that’s actually measured in numbers.
Subtypes:
Discrete Data: whole numbers only - number of hospital visits, for example.
Continuous Data: any value within a range - cholesterol level, for instance.
For our ICU group: age, height, and white blood cell count are all numerical.
Why Does This Actually Matter?
Because the type of data dictates the math. Some tests only work on numerical data; others are built for categorical data. You wouldn’t calculate an “average” blood type - A, B, AB, and O aren’t points on a number line, they’re categories.
Key Takeaways
Descriptive statistics summarizes what you have. Inferential statistics uses it to predict and test.
Categorical data groups observations; numerical data measures them.
Knowing your data type first is what tells you which statistical tools you’re even allowed to use.
Up Next:
Now that you can classify your data, we’ll cover the tools for summarizing it: measures of central tendency (mean, median, mode) and variability (range, variance, standard deviation), the essentials for making sense of any medical dataset.