[Guide] What to look for in a sample size calculator. This guide examines the most important factors you need to know when deciding what software to use for this important part of your study.

Why do I need a sample size calculator?

Using a sample size calculator to determine the appropriate sample size is a crucial step in study design. A study that has a sample size which is too small may produce inconclusive results and be considered unethical. Similarly, a study that has a sample size which is too large will waste scarce resources and could expose more participants than necessary to any related risk.

So what do you look for in a sample size calculator? Below we have listed the most important factors you need to know when deciding what software to use for this important part of your study.

Core Functionality

Different Clinical Trial Objectives:

Equivalence
Used when a trial is aiming to show that a new treatment is equivalent (within some margin) to a pre-existing treatment.

Non-inferiority
Used when a trial is aiming to show that a new treatment is not worse than an existing treatment by some pre-specified margin.

Superiority by a margin
Used when a trial is aiming to show that a new treatment is better than an existing treatment by some pre-specified margin.

Sample Sizes For Different Types of Data:

Normal
The data is continuous and assumed to follow a bell curve, centred at the mean and with the spread determined by the standard deviation.

Binary
The responses can only take on two possible states - usually related to success/ failure or yes/no. The data is summarised using proportions.

Ordinal
The data is discrete and takes on natural, ordered categories where the distance between each category is not known.

Clinical Trial Designs

Randomized Clinical Trials
Subjects are randomized to a treatment or control, such as a placebo.

Sample Size Statement for Regulatory Submission
Normal
This communicates the method and parameters used to calculate the sample size needed for the clinical trial.

Standard Functionality

Different Clinical Trial Objectives

Precision Based
Used for calculating confidence intervals, e.g

Bioequivalence
Used to show a new treatment has the same bioavailability as a standard treatment, within a given margin.

Sample Sizes for Different Types of Data

Survival
Methods for time-to-event data, often including hazard rates.

Rates & Counts
The frequency of events occurring over a period of time, common distributions include negative binomial and poisson.

Clinical Trial Designs

Cross-over
Giving patients a sequence of different treatments or dosages with a washout period between. This reduces the number of patients needed and the variability.

Parallel Group
Non-crossover study where different groups are given different treatments or
dosages.

Cluster Randomized
Clusters of subjects, such as doctors clinics, are randomized to each treatment such that all subjects in the same cluster receive the same treatment.

N-of-1 (Personalized Medicine)
Used when a trial is aiming to show that a new treatment is not worse than an existing treatment by some pre-specified margin.

Reporting
Detailed reporting

Advanced Functionality

Baselines/Covariates
Some methods, such as odds ratio tests, can be adjusted for multiple covariates.

Sensitivity Analysis
Analyzing the dependence of the sample size or power on specific parameters in the study.

Bayesian Sample Size
Using prior knowledge about a parameter to better estimate the probability of success of a trial.

Adaptive Designs for Efficacy
Sample size can be re-estimated in an adaptive trial depending on the probability of finding for efficacy.

Assessing Futility
Provision can be made in a study to stop for futility, if there is early evidence that the study will not find for efficacy.

Sample Size Re-Estimation
Sample size can be updated during the study to account for a changed variability or treatment effect.