Each week, we publish layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
The article featured today is from Statistics in Medicine, with the full article now available to read here.
, , , , , . Novel application of a discrete time-to-event model for randomized oral immunotherapy clinical trials with repeat food challenges. Statistics in Medicine. 2021; 1– 14. https://doi.org/10.1002/sim.9019
Food specific oral immunotherapy (OIT), which involves a food allergic-individual ingesting small amounts of an allergen in increasing dose amounts over a set period of time, has shown promise as an effective treatment in reducing food allergic reactions. The gold standard for assessing efficacy of an OIT-based treatment is a double-blind, randomized, placebo-controlled trial where participants undergo multiple oral food challenges over time throughout the course of the study. Typically, if a participant reacts during a food challenge, the participants will no longer partake in subsequent food challenges for the study. Due to this structure, some participants contribute more data than others (those who do not react at a given challenge will go on to contribute additional data on subsequent challenges whereas others who react at a given challenge will no longer contribute additional data on challenges).
This paper proposes a novel statistical method that defines a “dose-time” outcome by incorporating information on whether an allergic individual is able to complete a challenge without experiencing a reaction to the allergen, the dose at which a reaction occurred among those who experience reaction, and the timepoint in the study at which the challenge occurred. This “dose time” outcome acknowledges the following central tenets: 1) participants who pass more food challenges have better outcomes, 2) if a participant does have a reaction during an oral food challenge, it is considered better if they were able to tolerate more of their allergen, and 3) passing the next food challenge is more important than the dose at which a participant reacts.
Prior to this new method, statistical analysis of repeat oral food challenges involved looking at the proportion of participants who passed their challenge at a specific time point (i.e. what proportion of participants passed the 1-year challenge?). However, neither the dose at which they had a reaction nor the time point within the study course were used in the analysis. Thus, at six months, a participant who had a reaction at 350mg was treated the same as a participant who had a reaction at 1000mg. Further, individual trajectories were not incorporated into analyses where a participant followed for 1 year who reacted at 1000mg is different from one who reacted at the end of the 2-year study at 1000mg.
Overall, this newly proposed approach to analysis of repeated food challenges better leverages the data already collected from clinical trials designed to address the question of desensitization vs sustained unresponsiveness for treatments of food allergy. By utilizing more information and positioning food allergy data within the statistical framework of a “survival analysis”, this new approach can better detect meaningful differences between treatment arms.