2025 – PAGE 390 – STATISTICS

INCIDENCE

This refers to the number of new cases in a population over a given period of time. For example, if there were 30 new cases in a population of 2000 people over two years, the incidence would equal 7.5 cases per 1000 person-years (30 new cases divided by 4000 person-years).

PEARL: Person-years is the number of years that a person was studied. For example, if you follow 1 person for 2 years, that equals 2 person-years. If you follow 7 people for 3 years, that equals 21 person-years.

SAMPLE SIZE

This is the number of subjects being studied. A larger sample size will increase statistical power and increase the ability of a study to detect adverse events.

NUMBER NEEDED TO TREAT (NNT)

This is the number of patients that must be treated with the “better” treatment in order to prevent one bad outcome. The NNT depends on the specific outcome under consideration. For example, if the incidence of stroke in sickle cell patients is 0.15% with Treatment A (less strokes means better treatment) and 0.2% (more strokes means worse treatment) with Treatment B, then treating 2000 patients with A results in 3 strokes (0.0015 x 2000 = 3) and treating 2000 patients with B results in 4 strokes (0.0020 x 2000 = 4). Therefore, the NNT for preventing one stroke in one year is 2000. There is a reduction in risk of stroke when A is used instead of B. The Absolute Risk Reduction (ARR) in this example is 0.0020 − 0.0015 = 0.0005. To obtain the NNT, simply divide 1 by the ARR. So, 1/0.0005 = 2000.

RELATIVE RISK

This topic is covered under Cohort Studies.

ODDS RATIO

This topic is covered under Case Control Studies

KAPLAN MEIER CURVE (AKA KAPLAN-MEIER ESTIMATOR)

The Kaplan-Meier Estimator (used to create Kaplan-Meier curves) is used to estimate survival. The X-axis will always have a time component, and the Y-axis will typically have “Percent Survival.” The estimate can be useful to examine the probability of death, the effectiveness of a particular treatment, and recovery rates.

IMAGE: www.pbrlinks.com/kpcurve

PEARL: The hazard ratio is a low yield calculation that represents the instantaneous risk over the study time period. It differs from relative risk and odds ratio as these are cumulative over an entire study.

VALIDITY HIERARCHY

This refers to the relative authority, or “validity,” of the various types of study designs and types. There is an order of highest to lowest validity that is generally accepted. Systematic Review and Meta-Analysis >> Randomized Controlled Trials >> Cohort Studies >> Case-Controlled Studies >> Cross-Sectional Studies >> Case Studies. These topics will be discussed in order of highest validity to lowest validity.