Validated Web Tool Accurately Predicts Peanut Allergy in Children

An online calculator can predict peanut allergic reactions with high accuracy, a new study found.1

This Peanut Allergy Prediction Web Tool can speed allergy management, calculating the probability of an individual’s allergic reaction based on skin prick test, Ara h 2-sIgE, or basophil activation test results. Oral food challenges can be time-consuming, taking about 3 – 4 hours to complete, and this online tool could be a way for allergists to diagnose peanut allergies without all that waiting.2

“Currently, the suboptimal accuracy of routinely used tests leads to long waiting lists and often unnecessary allergen avoidance, that can increase the risk of peanut allergy,” wrote investigators, led by Alexandra F. Santos, MD, PhD, from Evelina London Children’s Hospital and St. Thomas’ Hospital in London.1

Santos and colleagues conducted this study to validate the diagnostic performance of the Peanut Allergy Prediction web tool in 173 children and to assess its ability to distinguish peanut-allergic vs peanut-sensitized tolerant cases. The team had previously developed the tool

The team collected data on demographics, clinical history, and results for skin prick test, Ara h 2-sIgE, basophil activation test, and oral food challenges. The study included 129 children with a confirmed peanut allergy and 44 with peanut sensitization, making it a 3:1 ratio of peanut-allergic to peanut-sensitized cases.

Investigators found the tool was excellent at distinguishing between peanut-allergic and peanut-sensitized patients. The basophil activation test results were the superior single test performer, with an area under the curve of 0.945. Ara h 2-sIgE was the second best (AUC, 0.879). The skin prick test, at last, had an AUC of 0.861.

Despite the basophil activation test being superior on its own, the 2-test combination of Ara h 2-sIgE and basophil activation test (AUC, 0.959) was the best at discriminating between allergy and sensitization cases and diagnostic accuracy. The 3-test combination, computing skin prick test, Ara h 2-sIgE, and basophil activation test results, was the third best diagnostic performance (AUC, 0.949).

The ROC analysis included cases with all 3 test results (n = 140) and showed consistent findings, with the skin prick test performing the least optimally (AUC, 0.853). The best combinations included the basophil activation test, specifically the combination of the basophil activation test and the Ara h 2-sIgE.

The sensitivity analysis demonstrated the high sensitivity (92.16%) but low specificity (63.16%) of the skin prick test, suggesting that only using this for the tool’s calculations could result in patients with peanut sensitizations receiving a peanut allergy diagnosis.

The basophil activation test results performed the best in sensitivity, specificity, and positive predictive value. Investigators wrote that its high positive predictive value indicates a high likelihood of detecting the presence of peanut allergy in the database.

The team noted several limitations, including an uneven number of peanut-allergic vs peanut-sensitized patients, missing test data reduced the number of cases in the analysis, and the low number of patients with severe reactions prevented the assessment of the tool’s ability to predict reaction severity. They wrote that further research with more cases is needed to expand the tool’s diagnostic accuracy.

“This study validates the online tool and supports its use with single, double, or triple test results to calculate reaction probabilities as an aid to diagnosis due to their effectiveness in accurately classifying allergic and sensitized tolerant cases,” investigators wrote. “The successful validation of the online prediction tool so far instills confidence in its ability to detect peanut allergy and discriminate between allergy and clinically irrelevant sensitization. Given this initial validation of the model’s performance, there is optimism for its clinical use in diagnosing peanut allergy in children.”

References

1. Mears H, Sanghera SA, Bahnson HT, et al. Open access online calculator predicts peanut allergic reactions with high accuracy. J Allergy Clin Immunol Pract. Published online May 14, 2025. doi:10.1016/j.jaip.2025.05.016

2. Food Challenge Test. Cleveland Clinic. https://my.clevelandclinic.org/health/diagnostics/9544-food-challenge-test. Accessed May 22, 2025.