Before any drug or treatment makes it to the market, it must go through a clinical trial. It is typically a series of tests that evaluate the effect of the drug on humans and determine the efficacy of the drug. But to ensure you get accurate results, a clinical trial must be planned in a scientific method. It is usually conducted after defining the core elements of the study including the research objective, statistical methods to evaluate the objective, and sample size estimation. Although every aspect of a clinical trial is crucial, sample size estimation is very important to get the right number of samples into the treatment.
In this article, we will discuss sample size estimation, the role it plays in clinical trials, and some strategies to reduce sample size for the feasibility of your study.
What is Sample Size?
The sample size is the number of individuals you need to study or observe for your research. If you are conducting a study to determine how many people commute on public transport at your workplace, and you want to conduct research with 50 employees, then 50 becomes your sample size.
What is Sample size estimation in clinical research?
In clinical research, sample size estimation is the process of choosing the right number of subjects so that you get statistically valid results. Usually, researchers pick a sample that closely represents the entire relevant population. The goal here is to pick a sample size that is large enough to produce reliable results and small enough to be cost-effective and manageable.
Biostatisticians play a critical role in clinical research, especially in determining the right sample size considering the type of study and its objectives. Sample size should be estimated carefully since studies are likely to fail even with large treatment effects because of miscalculated sample size.
Importance of sample size estimation in clinical trials
In a clinical trial, you test your drug or treatment against either an existing treatment or placebo. Determining an ideal sample size plays a crucial role in the outcome of a clinical trial. There are usually two scenarios while determining sample size, which will lead to a failed trial.
Sample size underestimation
Sample size underestimation is a scenario where you conduct a trial with an inadequate sample or number of subjects. In such a study, it will be difficult to identify the precise difference between the two interventions, which makes the study unethical. Also, as the sample size is too small to represent the entire target population, it cannot be generalized.
Sample size overestimation
In sample size overestimation, you select a sample size that is too large to logically identify the difference between the two treatments. This is because, with a large sample size even a small difference – which is considered clinically insignificant – will appear statistically significant. In addition, there is an ethical question of subjecting such a large group to risky treatment. For researchers, it also means a loss of time, effort, and resources in carrying out a trial.
Factors that affect sample size estimation in clinical trials
Accurate sample size estimation is derived statistically considering the study design parameters and hypotheses. Let’s take a look at different factors that need to be considered to determine sample size.
Null Hypotheses
In null hypotheses (H0), you conduct a trial to reject the new intervention against the existing one. It is basically set up to be rejected, by stating the opposite of what an investigator expects.
Alternative Hypotheses
Alternative hypotheses (H1) are the one where you confirm that the new treatment is better than the existing one. Its statement confirms the expectations of an investigator.
Type I error
Type I error is a false positive, where we reject the null hypothesis. The Type I error rate, i.e. the probability of rejecting a null hypothesis when it is actually true is known as the significance level. It is denoted by α and assumed that a 5% probability of making a type I error is acceptable.
Type II error
Type II error is a false negative, where we don’t reject a false null hypothesis. The type II error rate of failing to reject a false null hypothesis is denoted by β. The Power of a test is, hence, given as 1-β. Traditionally, power is taken as 80% – 90% when calculating the sample size.
Effect size is typically defined as the difference between the rate of an event in treatment group against the rate in control group. It gives you the practical significance of your clinical study, and it relies entirely on the data produced during the trial.
Reducing sample size for feasible clinical trials
Usually, determining the right sample size for clinical trials results in a large group of population. This presents two primary problems:
- Resources: Researchers will have to invest additional resources and cost to manage a large sample size.
- Ethical conundrum: Enrolling people ambitiously in a risky intervention could pose an ethical challenge.
Investigators can address these challenges by reducing sample size albeit with compromises like loss of power. However, with the help of biostatisticians, you can reduce sample size by not affecting the study objective to be analysed, while optimizing the resources.
Below is a list of strategies you can employ to reduce your sample size estimation:
- By reducing statistical power to 80%, you can reduce the sample size since such a compromise is not going to affect the data quality.
- Using continuous variables in your study gives you the option to have smaller sample sizes than dichotomous variables.
- Identifying a subject population that is more homogenous will reduce variability in outcome. However, this will impact the generalizability of the study.
- Improving patient experience to reduce dropout rates can help you optimize sample size. Through an intensive follow-up, you can drastically reduce the dropout rate.
- Expanding the minimal detectable difference can reduce your sample size but will be dependent on clinical experience or literature review.
- You can reduce sample size by increasing the event rate through exercises like the surrogate outcome, composite outcome, and an expanded follow-up period.
Sample size is key for valid clinical trials
For any clinical trial, rightly estimating sample size is one of the critical tasks that will require you to consider a whole list of factors. We discussed these factors very briefly to give you a higher-level understanding of sample size estimation and its role in clinical trials.
Biostatistics plays a critical role in estimating sample size, and it is important to have an experienced biostatistician determine this vital study parameter. At IQA, we have a team of very qualified and experienced biostatisticians who have worked across therapeutic areas.
To explore our biometrics services, reach out to us at hello@inductivequotient.com.
9 Comments
See all comments