Varying data analysis in ALS trials can muddle results, study finds

The statistical methods scientists use to analyze outcomes from clinical trials in amyotrophic lateral sclerosis (ALS) vary widely, and these differences may make some trials prone to generating misleading results, a study found.

The researchers focused on analysis of the ALS Functional Rating Scale-Revised (ALSFRS-R), a measure of functional ability often used as a primary outcome in ALS clinical trials.

“We demonstrate that even randomized trials, that are otherwise well-designed for sample size, duration, and conduct, can yield biased or misleading conclusions when suboptimal statistical analysis strategies are applied,” the researchers wrote. “These inconsistencies can mask true treatment effects or falsely suggest benefit, thereby undermining the translation of promising drugs into successful clinical trials or erroneously advancing ineffective therapies.”

The scientists called for efforts to develop recommendations to ensure trials are using similar statistical approaches to analyze the disability scale in ALS and similar scales in other neurodegenerative diseases.

The study, “Heterogeneity in the Analysis of the ALSFRS-R in ALS Clinical Trials and its Effect on the Validity and Precision of Trial Conclusions,” was published in Neurology.

Recommended Reading

June 5, 2026 News by Marisa Horak, MS

New ALS research effort aims to speed search for treatments

Analyzing trial analysis

To track ALS progression, researchers and clinicians usually rely on the ALSFRS-R, which measures how well an individual can do different activities that are impacted by ALS, such as walking, talking, and breathing.

Someone who can do the activity perfectly scores a 4, while someone who can’t do it at all scores a 0. The individual scores are then added for a total possible score of 48. In ALS, scores on this scale decrease over time as the disease progresses and patients lose functionality.

Clinical trials testing ALS treatments usually rely on the ALSFRS-R as a primary measure of effectiveness. But from a mathematical perspective, there are several ways to analyze data from a numerical tool like the ALSFRS-R.

For example, scientists may choose to perform cross-sectional analyses — that is, looking at scores at a single point in time — or they may analyze data longitudinally, comparing how scores evolve over weeks or years of follow-up, which has much greater statistical power to detect reliable results.

Researchers may also choose to control the data for each patient’s starting ALSFRS-R score or for other differences between patients that are known to affect how quickly ALS progresses, but the factors that are included in the analyses vary widely.

Another issue is that not all patients have complete data, so researchers must decide what to do when scores are missing — and different choices can yield different results.

The team of scientists in the Netherlands and Canada dug into data from 45 large ALS clinical trials to determine which statistical analyses were used to analyze ALSFRS-R scores. The “aim was to characterize current analytical practices for the ALSFRS-R in randomized controlled trials and to assess how these choices influence the validity and precision of trial conclusions,” the researchers wrote.

The researchers found substantial variability in the analyses used: Across the 45 trials, 39 different statistical approaches were used. And more than half of the studies did not include all the ALSFRS-R measurements taken over the course of the trial, “resulting in suboptimal utilization of patient data and reduced statistical precision,” the scientists wrote.

“Despite the clear evidence that inconsistencies in statistical methods can influence trial conclusions and lead to discrepant results, our findings reveal the continued use of statistically invalid approaches in a substantial proportion of studies,” the team wrote, noting that invalid statistical analyses “complicate comparisons across trials, hinder regulatory decision-making, and lead to inefficient use of limited resources, funding, and time.”

To illustrate how these different statistical approaches can affect outcomes, the researchers applied the different approaches to the same dataset from various trials. They found that some statistical approaches tended to exaggerate the difference between the placebo and treatment groups, potentially leading to false-positive results.

In fact, they said, more than a third (38.9%) of the trials used statistical frameworks that “were at risk of increasing false-positive rates, potentially contributing to the erroneous advancement of ineffective treatments.”

Efforts to standardize the statistical approaches used to analyze outcomes in trials of ALS and other diseases are needed, the researchers said.

“In ALS, this process is already underway through a dedicated working group,” they wrote. “Considering the considerable time, resources, and commitment of trial participants, it is essential to maximize the value of collected data through rigorous statistical methods.”