Statistics for Nurses

Nurses use statistics every day — from reading lab results to evaluating whether a new treatment is supported by evidence. Understanding the numbers behind clinical practice is not optional; it is the foundation of safe, effective patient care. Whether you are interpreting a blood pressure reading, assessing the reliability of a diagnostic test, or reading a research article about a new medication, statistical literacy helps you make better decisions at the bedside.

This page covers the statistical concepts that matter most in clinical nursing practice. It builds on the ideas from Mean, Median, and Mode and Normal Distribution.

Interpreting Vital Signs Data

Every set of vital signs you record is a data point. The “normal ranges” that guide your clinical judgment are statistical constructs based on large populations.

When we say a value is “normal,” we mean it falls within approximately 2 standard deviations of the population mean. This range captures about 95% of healthy individuals.

Example: Heart Rate Reading

A patient presents with a resting heart rate of 100 bpm. Is this concerning?

If the population mean resting heart rate for healthy adults is about 80 bpm with SD of approximately 10 bpm, then 2 standard deviations above the mean is:

$80 + 2(10) = 100 \text{ bpm}$

A reading of 100 bpm sits right at the boundary — it is approximately 2 standard deviations above the mean. While a single reading at this level does not necessarily indicate a problem, it is a clinical flag that warrants follow-up. This is exactly how statistical thinking informs triage decisions.

Understanding Lab Result Ranges

Reference ranges on lab reports are typically the middle 95% of results from a healthy reference population. Mathematically, they represent:

$\text{Reference range} = \mu \pm 2\sigma$

where $\mu$ is the population mean and $\sigma$ is the standard deviation.

A critical implication: about 5% of perfectly healthy people will have a lab value outside the reference range. A value flagged as “high” or “low” is not automatically abnormal — it simply means the result falls in the outer tails of the healthy distribution.

Example: Hemoglobin Reference Range

The reference range for hemoglobin in adult females is typically 12.0 — 16.0 g/dL. If the population mean is 14.0 g/dL with SD of 1.0 g/dL:

$\mu - 2\sigma = 14.0 - 2(1.0) = 12.0$

$\mu + 2\sigma = 14.0 + 2(1.0) = 16.0$

A patient with hemoglobin of 11.8 g/dL is just below the reference range. This could indicate anemia — or the patient could be among the approximately 2.5% of healthy women whose hemoglobin naturally falls below 12.0 g/dL. Clinical context (symptoms, history, trends) matters as much as the number itself.

Key nursing takeaway: Never rely on a single lab value in isolation. Look at trends over time, clinical symptoms, and the full picture.

Sensitivity and Specificity

When evaluating diagnostic tests, two metrics tell you how well the test performs:

• Sensitivity = $P(\text{positive test} \mid \text{disease present})$ = true positive rate. A test with high sensitivity is good at catching disease. Few people with the disease will be missed (few false negatives).

• Specificity = $P(\text{negative test} \mid \text{no disease})$ = true negative rate. A test with high specificity is good at ruling out disease. Few healthy people will be falsely flagged (few false positives).

A helpful mnemonic: SnNOut (Sensitive test, Negative result, rules Out) and SpPIn (Specific test, Positive result, rules In).

Example 1: Rapid Strep Test

A rapid strep test has sensitivity 85% and specificity 98%. In a clinic population where 10% of patients actually have strep, what happens when you test 1,000 patients?

Step 1 — Set up the population.

• 1,000 patients total
• 100 actually have strep (10%)
• 900 do not have strep (90%)

Step 2 — Apply sensitivity to the diseased group.

$\text{True positives} = 100 \times 0.85 = 85$

$\text{False negatives} = 100 - 85 = 15$

Step 3 — Apply specificity to the healthy group.

$\text{True negatives} = 900 \times 0.98 = 882$

$\text{False positives} = 900 - 882 = 18$

Step 4 — Calculate predictive values.

$\text{PPV} = \frac{\text{True positives}}{\text{All positive tests}} = \frac{85}{85 + 18} = \frac{85}{103} \approx 0.825 = 82.5\%$

$\text{NPV} = \frac{\text{True negatives}}{\text{All negative tests}} = \frac{882}{882 + 15} = \frac{882}{897} \approx 0.983 = 98.3\%$

Interpretation: If a patient tests positive, there is an 82.5% chance they actually have strep. If a patient tests negative, there is a 98.3% chance they truly do not have strep. The high NPV means a negative result is very reassuring.

Nursing action: Because sensitivity is only 85%, about 15% of strep cases will be missed by the rapid test. This is why a negative rapid strep result is often followed by a throat culture (which has higher sensitivity).

Odds Ratios and Relative Risk

When reading research articles, you will encounter two measures that quantify the association between an exposure and an outcome:

Relative risk (used in cohort studies and RCTs):

$RR = \frac{P(\text{outcome} \mid \text{exposed})}{P(\text{outcome} \mid \text{unexposed})}$

Odds ratio (used in case-control studies):

$OR = \frac{a \times d}{b \times c}$

from a 2-by-2 table where $a$ = exposed with outcome, $b$ = exposed without outcome, $c$ = unexposed with outcome, $d$ = unexposed without outcome.

Example 2: Smoking and Heart Disease

A cohort study tracks smokers and non-smokers over 10 years:

Relative risk:

$P(\text{heart disease} \mid \text{smoker}) = \frac{40}{200} = 0.20$

$P(\text{heart disease} \mid \text{non-smoker}) = \frac{20}{300} \approx 0.0667$

$RR = \frac{0.20}{0.0667} \approx 3.0$

Smokers are 3 times more likely to develop heart disease than non-smokers.

Odds ratio:

$OR = \frac{a \times d}{b \times c} = \frac{40 \times 280}{160 \times 20} = \frac{11{,}200}{3{,}200} = 3.5$

The odds ratio (3.5) is slightly higher than the relative risk (3.0). This is typical — when the outcome is common (20% among smokers), the OR overestimates the RR. For rare outcomes, the two are approximately equal.

Reading Clinical Trial Results

Evidence-based practice requires nurses to understand the basics of randomized controlled trials (RCTs).

Key concepts:

• Randomization — patients are randomly assigned to treatment or control, minimizing bias
• Intention-to-treat analysis — all patients are analyzed in the group they were assigned to, even if they did not complete treatment
• Confidence intervals — a range of plausible values for the true treatment effect. If a 95% CI for a difference in means does not contain 0, the result is statistically significant at the 0.05 level.

Number Needed to Treat (NNT)

The NNT tells you how many patients you need to treat to prevent one bad outcome. It is calculated from the absolute risk reduction (ARR):

$ARR = P(\text{event in control group}) - P(\text{event in treatment group})$

$NNT = \frac{1}{ARR}$

Example 3: Antibiotic Prophylaxis

A study finds that a prophylactic antibiotic reduces surgical site infection from 12% to 4%.

$ARR = 0.12 - 0.04 = 0.08$

$NNT = \frac{1}{0.08} = 12.5$

Since you cannot treat half a patient, round up: you would need to treat 13 patients with the prophylactic antibiotic to prevent one surgical site infection. This is a clinically meaningful result that helps nurses and physicians weigh the benefits of the intervention against its costs and side effects.

Evaluating Evidence-Based Practice

Not all evidence is created equal. The hierarchy of evidence ranks study types by their reliability:

1. Systematic reviews and meta-analyses — combine results from multiple high-quality studies
2. Randomized controlled trials (RCTs) — gold standard for establishing cause and effect
3. Cohort studies — follow groups over time, but without randomization
4. Case-control studies — compare people with and without an outcome, looking back at exposures
5. Case reports and expert opinion — useful for generating hypotheses, but weakest evidence

When evaluating a study for clinical practice, ask:

• Was it randomized? (Reduces bias)
• Was the sample size large enough? (Increases statistical power)
• Are the confidence intervals narrow? (Increases precision)
• Is the NNT clinically meaningful? (Practical significance, not just statistical significance)
• Does it apply to your patient population? (Generalizability)

Practice Problems

Test your understanding with these problems. Click to reveal each answer.

Key Takeaways

• Normal ranges for vital signs and lab values represent the middle 95% of a healthy population (mean plus or minus 2 standard deviations) — about 5% of healthy people will fall outside these ranges
• Sensitivity measures how well a test catches disease; specificity measures how well it rules out disease. Both matter, and their clinical usefulness depends on disease prevalence.
• Relative risk compares event rates between groups; odds ratio is used when you cannot directly calculate rates (case-control studies)
• NNT (Number Needed to Treat) translates statistical results into practical clinical terms
• The hierarchy of evidence helps nurses evaluate which studies should carry the most weight in clinical decisions

Return to Statistics for more topics in this section.