New method improves RMST analysis for clinical and epidemiological studies

The limited average Survival Time (RMST) analysis technique was introduced in health care research about 25 years ago and has since been widely used in economics, engineering, companies and other professions.

In clinical environments, RMST is useful because it is a simple way to understand the average survival time-the duration that patients live after diagnosis or treatment and the factors that affect that time in a specified time frame.

In addition, in contrast to COX registration models and other popular models, trusts and comparisons that are made using RMST not on the proportion hazard that the chance of an event will be constant over time.

But there is a catch: RMST can test differences in the effect of a treatment between groups of baseline up to a time-the threshold but identifying the ideal threshold in clinical and epidemiological studies is difficult. This leads to results that are less statistically powerful than they could be. “

Gang Han, PhD, Professor Biostatistics, Texas A & M University School of Public Health

In order to take on this challenge, Han and colleagues in the academic world and industry have developed a new method that uses an existing mathematical tool-the reduced itemly exponentially model to determine the ideal or optimal threshold time in the limited average survival analysis when studying two groups.

“This is especially important in medical studies, because the chance of a specific event can change in the different stages of treatment,” said Matthew Lee Smith, PhD, professor in health behavior at the Texas A&M School of Public Health, which was involved in this study.

To determine the optimum threshold, the team calculated a threshold time of significant change points in hazard percentages and compared what they found with the greatest possible threshold time.

Their research paper, published in the American Journal of Epidemiologyshowed the benefits of the proposed method in multiple simulation studies and two real examples, a clinical study and an epidemiology study.

They used the new method to measure type 1 error percentages and statist capacity in simulations in which the sailing speed was constant for one group and was changed for another group. They compared the groups using the standard Logrank test and their new model.

“Our model performed the best,” said Marcia G. Ory, PhD, Regents and Distingughed Professor in the School of Public Health investigating evidence-based prevention methods. “That was also the case when we applied it to two Real-World scenarios.”

For both scenarios, traditional statistical analysis methods did not reveal remarkable differences between two treatments. However, when the new model was applied, the results for each scenario discovered that one treatment was clearly superior.

The first scenario compared two treatments for seven months for patients with non-smalllling lung cancer who had lower levels of an important biomarker. The second used a standard assessment to measure the time to purchase people with mild dementia who lived with care providers compared to those who did not live with care providers.

“These results are promising, and more research is needed that compares more than two groups and that uses multiple covariates, such as the age of the participants, ethnicity and socio -economic status,” said Han. “Nevertheless, we believe that this method could be more powerful based on these early results than all existing comparisons for two groups in the analysis of the results of time-to-event.”

Others involved in the study were the Epidemiology and Biostatistics Doctorate student Laura Hopkins, Raymond Carroll, PhD, Distinguished Professor in the Texas A&M Department of Statistics and external employees of Eli Lilly and Company & Het H. Lee.