At the conclusion of a clinical trial, researchers are left with a collection of data to interpret. In a study assessing an ALS drug, data such as participant survival, ALSFRS-r scores, and biomarkers, are usually analyzed by comparing results between the active drug and placebo groups. To determine whether a trial met its endpoints – the outcomes or events that are being evaluated during a trial – researchers will look to see if the group receiving active drug had significantly different results from the placebo group.

A crucial aspect of this analysis is ensuring that observed differences between the active drug and placebo groups, such as differences in survival or ALSFRS-r scores, are not due to random chance. The p-value serves as a statistical measure of this chance. A lower p-value indicates researchers can be more confident that observed differences, whether positive or negative, are likely due to the treatment rather than chance. ­

What is a “null hypothesis?”

Understanding p-values involves grasping the concepts of the "null hypothesis" and "alternative hypothesis." In a statistical study, a null hypothesis is the assumption that that there is no difference between two groups. Good examples to understand this are the many observational studies that have shown connections between ALS and military service.

  • Today, it is well established that veterans are nearly twice as likely than the general population to develop ALS.
  • If a researcher was designing a study to confirm this, the null hypothesis could be that there is no difference in prevalence of ALS for people in the military versus the general population.
  • The alternative hypothesis would be that ALS is more prevalent among people who have served in the military.

When conducting a study like this, researchers would gather data about ALS prevalence, separating people who have served in the military from the rest of the population and looking for any differences between the two groups. If a difference was found, they would then calculate the likelihood that their results were caused by random chance.

The p-value represents the chance, determined by these calculations, that the study’s results would have occurred if the null hypothesis was true. In this case, if the researchers saw an elevated risk with a p-value of p=0.05, it would mean that there is a 5% chance of seeing those results if, in reality, there was no increased rate of ALS in the miliary. Results that are very unlikely to be caused by chance, with a low p-value, are often referred to as “statistically significant.”

P-values in Clinical Trials

Like an observational population study, a clinical trial has a null and alternative hypothesis. In a trial meant to demonstrate a treatment’s efficacy, these could be summarized to “the drug did not have any effect on the participants” and “the drug did have an effect on participants.” When calculating the p-value of their results, the researchers would look at the chances that any differences between the on-drug and placebo groups would occur even if the drug had no effect. If the data for the on-drug group met the study’s endpoints with a low p-value, it could indicate that the drug was having a therapeutic effect.

What factors affect a clinical trial’s p-value?

When designing a clinical trial that will test a drug’s efficacy, it is important for researchers to make sure that they are recruiting the right participants to produce reliable data.
  • First, researchers must include enough people – the more people participate in a study, the lower the chance that any results might be caused by chance. For example, in an ALS study with 30 participants, having one person who naturally progresses faster in the active drug group could bring down the overall progression data of the entire group when these numbers are averaged. In a study with 1,000 participants, this one person would have much less effect on the overall data.
This is why phase 3 trials – which test drugs with the goal of generating data strong enough for regulatory approval – tend to enroll larger numbers of participants than earlier phase trials. Studies that do not have enough participants to generate reliable results are often said to be “underpowered.”

Researchers must also use carefully selected inclusion and exclusion criteria to account for random chance. These criteria determine who can – and who cannot – enroll in a trial. They are necessary to make sure that, in the case of an ALS trial, factors such as a participant’s level of progression, the medications or supplements they take, or their inability to participate in all aspects of the study do not influence the results.

Another element of clinical trial study design used to account for random chance is the randomization ratio.

  • This ratio represents the number of participants who will be assigned to the active drug group compared to those on placebo.
  • A trial with a 1:1 randomization ratio will have the same number of participants on active drug and placebo, while one with a 2:1 or 3:1 ratio has two times or three times as many on drug, respectively.

When determining this ratio, researchers must strike a careful balance. Particularly in a terminal disease like ALS, many people feel it is important to make sure as many participants as possible have access to an active treatment. However, it is also important to make sure there are enough participants in the placebo group to generate reliable data for comparison with the on-drug group.

Can we learn anything from studies with a high p-value?

Even if an analysis produces results with a p-value above the statistically significant  range – often around p=0.05 – it can still offer valuable insights that researchers can learn from. One common example of this is post-hoc analyses of clinical trial data. In a post-hoc analysis, researchers look at data from a failed trial to see if any subgroups of participants – such as those with a particular gene mutation – might have benefited from the treatment, even if most participants did not.

The results of these post-hoc analyses studies often have higher p-values, because they look at smaller samples of participants from the original study. These subgroups also may not be evenly distributed between the on-drug and placebo groups. While the results of a post-hoc analysis may often lack the reliability required to submit a drug for approval, they can point the way for researchers to design a new, more targeted trial.


Where can I learn more about ALS clinical trials?

The ALS Therapy Development Institute (ALS TDI) provides information about currently recruiting ALS trials as a service for our community. You can currently search for clinical trials that might interest you with our Clinical Trials Database. In early 2024, we plan to release a new tool, the ALS Trial Navigator, which will allow users to generate a customized list of ALS trials that meet their specific criteria.

To stay up to date on the ALS Trial Navigator, and see other updates about the research at ALS TDI, subscribe to our newsletter.

What to Do Next