Clinical trials are complicated and often the results can be even more so. Close monitoring of the interactions between patients, diseases, conditions and drugs for extended periods of time and the various questions posed in order to prove or disprove safety and efficacy of particular treatments result in an accurate yet highly complex mountain of data.
Clinical trials are conducted to answer a specific question about a treatment, usually related to safety and efficacy. This involves the development of a hypothesis, and then the running of clinical trials to collect data and to demonstrate the validity of the hypothesis using statistical analysis.
To most of us the data, and even the resulting analysis, are difficult to understand in real terms. The statistical analysis, while necessary to demonstrate the strength of evidence in favor of a hypothesis, also uses terms and tools that are challenging to patients, investors and the general public.
One of the most important values that can be assigned to findings by the process of statistical analysis is the p-value, or ‘probability’ value.
The p-value is a number between 0.00 and 1.0, and is used to demonstrate the strength of a conclusion drawn from clinical trial data. It enables analysts to assign a widely accepted numerical value to the strength of a statement or hypothesis. Essentially the p-value measures consistency between the results actually obtained in the trial and the “pure chance” explanation for those results.
To illustrate a clinical trial structure very simply; a question in regards to a treatment may be “does x work as a treatment for y?”. From this, both affirmative and negative hypothesis would be posed as answers, ‘no x does not work’ and ‘yes x does work’. A clinical trial would then be run and data generated and collated. Each hypothesis would be tested against this data set; this demonstration uses statistical analysis and is expressed in terms of probability using a p-value, which communicates the strength of a finding present in the data set.
For example in our Phase III interim trial results for erythropoietic protoporphyria (EPP), released in January this year we announced:
The maximum severity of phototoxic reactions was significantly reduced by afamelanotide treatment compared with placebo (p<0.001)
What this means is that there is very little chance (less than 0.1%, or 1 chance in 1000) of the same reduction in severity of phototoxic reactions occurring by chance. This makes our finding, and this statement of strong statistical significance.
A statement and corresponding p-value are considered of strong significance if the probability of the same reaction occurring randomly or by chance is less than 5%, corresponding to a p-value of p<0.05
While complex and confusing, especially when viewed within results and company announcements, p-values give a valuable indication of the strength of a statement drawn from clinical trial data. Once understood, p-values are an excellent tool for quickly ascertaining the significance of a statement made about treatment safety and efficacy.
