Predicting first-time anaphylaxis in the elderly using stacked machine learning and population registries

Aug 4, 2026

The study develops and evaluates a stacked machine learning approach to predict first-time anaphylaxis among older adults using population registry information. It compares multiple modeling strategies and examines performance using classification metrics and calibration assessment. Feature contributions are analyzed to identify factors that most strongly support risk discrimination. The results quantify predictive accuracy on real-world data, highlighting the potential for registry-linked machine learning to support earlier identification of high-risk individuals in clinical and public-health settings.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Methods

Data sources and population registries

Stacked machine learning approach

Results

Model performance and evaluation

Feature importance and risk factors

Discussion

Clinical and public-health implications

􀀘Y􀀁E O􀀌􀀇􀀕􀀇􀀈􀀖􀀚 R􀀂􀀍􀀂􀀖􀀌􀀓􀀔 􀀁UBLISHE􀀝 06 N􀀐􀀋􀀂􀀃b􀀂􀀌 2025 􀀝OI 10 3389/􀀑􀀖􀀚􀀕􀀏 2025 1655662 􀀁􀀂􀀃􀀄􀀁􀀂 􀀆􀀇 􀀁􀀂􀀃 􀀅􀀆􀀇􀀈􀀉 􀀊􀀈􀀇􀀋􀀂􀀌􀀍􀀇􀀎􀀏 􀀐􀀑 􀀒􀀇􀀓􀀔􀀇􀀕􀀖􀀈􀀉 􀀊􀀈􀀇􀀎􀀂􀀗 􀀘􀀎􀀖􀀎􀀂􀀍 􀀈􀀁􀀉􀀃􀀁􀀊􀀁􀀂 􀀆􀀇 􀀘􀀖􀀌􀀖 􀀒􀀖􀀈􀀎􀀇􀀉 􀀊􀀈􀀇􀀋􀀂􀀌􀀍􀀇􀀎􀀏 􀀐􀀑 􀀒􀀂􀀍􀀍􀀇􀀈􀀖􀀉 􀀙􀀎􀀖􀀚􀀏􀀁􀀔􀀖􀀌􀀚􀀂􀀍 􀀘􀀓􀀔􀀛􀀚􀀂􀀌􀀉 􀀊􀀈􀀇􀀋􀀂􀀌􀀍􀀇􀀎􀀏 􀀐􀀑 􀀒􀀇􀀓􀀔􀀇􀀕􀀖􀀈􀀉 􀀊􀀈􀀇􀀎􀀂􀀗 􀀘􀀎􀀖􀀎􀀂􀀍 􀀜􀀋􀀌􀀈􀀈􀀁􀀍􀀎􀀌􀀏􀀂􀀁􀀏􀀋􀀁 􀀝􀀐􀀈􀀇 􀀒􀀐􀀌􀀖 􀀎􀀃􀀐􀀌􀀖􀀞􀀛􀀇􀀓 􀀂􀀍 􀀈􀀁􀀋􀀁􀀃􀀉􀀁􀀂 28 J􀀛􀀈􀀂 2025 􀀐􀀋􀀋􀀁􀀎􀀄􀀁􀀂 10 O􀀓􀀎􀀐b􀀂􀀌 2025 􀀎􀀑􀀆􀀒􀀃􀀍􀀓􀀁􀀂 06 N􀀐􀀋􀀂􀀃b􀀂􀀌 2025 􀀋􀀃􀀄􀀐􀀄􀀃􀀌􀀏 􀀒􀀐􀀌􀀖 􀀝􀀉 R􀀐􀀓􀀔􀀂 D 􀀖􀀈􀀗 􀀒􀀛ñ􀀐z-􀀁􀀖􀀈􀀐 R (2025) P􀀌􀀂􀀗􀀇􀀓􀀎􀀇􀀈􀀕 ﬁ􀀌􀀍􀀎-􀀎􀀇􀀃􀀂 􀀖􀀈􀀖p􀀔􀀏􀀚􀀖x􀀇􀀍 􀀇􀀈 􀀎􀀔􀀂 􀀂􀀚􀀗􀀂􀀌􀀚􀀏􀀛􀀍􀀇􀀈􀀕 􀀍􀀎􀀖􀀓􀀆􀀂􀀗 􀀃􀀖􀀓􀀔􀀇􀀈􀀂 􀀚􀀂􀀖􀀌􀀈􀀇􀀈􀀕 􀀖􀀈􀀗 p􀀐p􀀛􀀚􀀖􀀎􀀇􀀐􀀈 􀀌􀀂􀀕􀀇􀀍􀀎􀀂􀀌􀀍 F􀀌􀀐􀀈􀀎 􀀅􀀚􀀚􀀂􀀌􀀕􀀏 6:1655662 􀀗􀀐􀀇: 10 3389/􀀑􀀖􀀚􀀕􀀏 2025 1655662 􀀋􀀌􀀎􀀇􀀈􀀃􀀔􀀓􀀄 © 2025 􀀒􀀐􀀌􀀖􀀉 R􀀐􀀓􀀔􀀂 􀀖􀀈􀀗 􀀒􀀛ñ􀀐z-􀀁􀀖􀀈􀀐 􀀝􀀔􀀇􀀍 􀀇􀀍􀀖􀀈 􀀐p􀀂􀀈-􀀖􀀓􀀓􀀂􀀍􀀍 􀀖􀀌􀀎􀀇􀀓􀀚􀀂 􀀗􀀇􀀍􀀎􀀌􀀇b􀀛􀀎􀀂􀀗 􀀛􀀈􀀗􀀂􀀌 􀀎􀀔􀀂􀀎􀀂􀀌􀀃􀀍 􀀐􀀑 􀀎􀀔􀀂 􀀁􀀌􀀂􀀖􀀎􀀇􀀋􀀂 􀀁􀀐􀀃􀀃􀀐􀀈􀀍 􀀅􀀎􀀎􀀌􀀇b􀀛􀀎􀀇􀀐􀀈 L􀀇􀀓􀀂􀀈􀀍􀀂 (􀀁􀀁 BY) 􀀝􀀔􀀂 􀀛􀀍􀀂􀀉 􀀗􀀇􀀍􀀎􀀌􀀇b􀀛􀀎􀀇􀀐􀀈 􀀐􀀌􀀌􀀂p􀀌􀀐􀀗􀀛􀀓􀀎􀀇􀀐􀀈 􀀇􀀈 􀀐􀀎􀀔􀀂􀀌 􀀑􀀐􀀌􀀛􀀃􀀍 􀀇􀀍 p􀀂􀀌􀀃􀀇􀀎􀀎􀀂􀀗􀀉 p􀀌􀀐􀀋􀀇􀀗􀀂􀀗 􀀎􀀔􀀂 􀀐􀀌􀀇􀀕􀀇􀀈􀀖􀀚 􀀖􀀛􀀎􀀔􀀐􀀌(􀀍) 􀀖􀀈􀀗 􀀎􀀔􀀂􀀓􀀐p􀀏􀀌􀀇􀀕􀀔􀀎 􀀐w􀀈􀀂􀀌(􀀍) 􀀖􀀌􀀂 􀀓􀀌􀀂􀀗􀀇􀀎􀀂􀀗 􀀖􀀈􀀗 􀀎􀀔􀀖􀀎 􀀎􀀔􀀂􀀐􀀌􀀇􀀕􀀇􀀈􀀖􀀚 p􀀛b􀀚􀀇􀀓􀀖􀀎􀀇􀀐􀀈 􀀇􀀈 􀀎􀀔􀀇􀀍 j􀀐􀀛􀀌􀀈􀀖􀀚 􀀇􀀍 􀀓􀀇􀀎􀀂􀀗􀀉 􀀇􀀈􀀖􀀓􀀓􀀐􀀌􀀗􀀖􀀈􀀓􀀂 w􀀇􀀎􀀔 􀀖􀀓􀀓􀀂p􀀎􀀂􀀗 􀀖􀀓􀀖􀀗􀀂􀀃􀀇􀀓 p􀀌􀀖􀀓􀀎􀀇􀀓􀀂 N􀀐 􀀛􀀍􀀂􀀉 􀀗􀀇􀀍􀀎􀀌􀀇b􀀛􀀎􀀇􀀐􀀈 􀀐􀀌 􀀌􀀂p􀀌􀀐􀀗􀀛􀀓􀀎􀀇􀀐􀀈 􀀇􀀍 p􀀂􀀌􀀃􀀇􀀎􀀎􀀂􀀗 w􀀔􀀇􀀓􀀔 􀀗􀀐􀀂􀀍 􀀈􀀐􀀎 􀀓􀀐􀀃p􀀚􀀏 w􀀇􀀎􀀔􀀎􀀔􀀂􀀍􀀂 􀀎􀀂􀀌􀀃􀀍 􀀁􀀂􀀃􀀄􀀅􀀆􀀇􀀅􀀈􀀉 􀀋􀀂􀀌􀀇􀀍􀀇􀀅􀀎􀀃 􀀏􀀈􀀏􀀐􀀑􀀒􀀓􀀏􀀔􀀅􀀌􀀅􀀈 􀀇􀀑􀀃 􀀃􀀓􀀄􀀃􀀂􀀓􀀒 􀀕􀀌􀀅􀀈􀀉 􀀌􀀇􀀏􀀆􀀖􀀃􀀄􀀎􀀏􀀆􀀑􀀅􀀈􀀃 􀀓􀀃􀀏􀀂􀀈􀀅􀀈􀀉 􀀏􀀈􀀄􀀐􀀗􀀐􀀕􀀓􀀏􀀇􀀅􀀗􀀈 􀀂􀀃􀀉􀀅􀀌􀀇􀀃􀀂􀀌 􀀘􀀗􀀈􀀅 􀀙􀀗􀀂􀀏 􀀚􀀛􀀜 􀀝􀀏􀀞􀀅􀀄 􀀗􀀆􀀑􀀃􀀚 􀀏􀀈􀀄 􀀗􀀌􀀏 􀀙􀀕ñ􀀗z􀀍C􀀏􀀈􀀗 2􀀜3􀀜4 1 R􀀂􀀍􀀂􀀖􀀌􀀓􀀔 􀀙􀀈􀀍􀀎􀀇􀀎􀀛􀀎􀀂 􀀑􀀐􀀌 E􀀋􀀖􀀚􀀛􀀖􀀎􀀇􀀐􀀈 􀀖􀀈􀀗 P􀀛b􀀚􀀇􀀓 P􀀐􀀚􀀇􀀓􀀇􀀂􀀍 (􀀙R􀀅PP)􀀉 􀀊􀀈􀀇􀀋􀀂􀀌􀀍􀀇􀀎􀀖􀀎 􀀙􀀈􀀎􀀂􀀌􀀈􀀖􀀓􀀇􀀐􀀈􀀖􀀚 􀀗􀀂 􀀁􀀖􀀎􀀖􀀚􀀛􀀈􀀏􀀖(􀀊􀀙􀀁)􀀉 B􀀖􀀌􀀓􀀂􀀚􀀐􀀈􀀖􀀉 􀀘p􀀖􀀇􀀈􀀉 2􀀅􀀚􀀚􀀂􀀌􀀕􀀏 D􀀂p􀀖􀀌􀀎􀀃􀀂􀀈􀀎􀀉 H􀀐􀀍p􀀇􀀎􀀖􀀚 􀀁􀀚􀀇􀀈􀀇􀀓􀀉 B􀀖􀀌􀀓􀀂􀀚􀀐􀀈􀀖􀀉 􀀘p􀀖􀀇􀀈􀀉 3􀀒􀀂􀀝R􀀙 2􀀅—􀀙􀀈􀀍􀀎􀀇􀀎􀀛􀀎􀀗’􀀙􀀈􀀋􀀂􀀍􀀎􀀇􀀕􀀖􀀓􀀇􀀐􀀈􀀍 B􀀇􀀐􀀃è􀀗􀀇q􀀛􀀂􀀍 􀀅􀀛􀀕􀀛􀀍􀀎 P􀀇 􀀙 􀀘􀀛􀀈􀀏􀀂􀀌 (􀀙D􀀙B􀀅P􀀘)􀀉 B􀀖􀀌􀀓􀀂􀀚􀀐􀀈􀀖􀀉 􀀘p􀀖􀀇􀀈􀀉 4R􀀙􀀁OR􀀘—RE􀀙-􀀙􀀈􀀍􀀎􀀇􀀎􀀛􀀎􀀐􀀗􀀂 􀀘􀀖􀀚􀀛􀀗 􀀁􀀖􀀌􀀚􀀐􀀍 􀀙􀀙􀀙􀀉 􀀒􀀖􀀗􀀌􀀇􀀗􀀉 􀀘p􀀖􀀇􀀈 􀀆􀀕􀀖􀀗􀀘􀀙􀀚􀀛􀀜􀀝􀀞 􀀅􀀈􀀖p􀀔􀀏􀀚􀀖x􀀇􀀍 􀀇􀀍 􀀖 􀀍􀀂􀀋􀀂􀀌􀀂􀀉 p􀀐􀀎􀀂􀀈􀀎􀀇􀀖􀀚􀀚􀀏 􀀚􀀇􀀑􀀂-􀀎􀀔􀀌􀀂􀀖􀀎􀀂􀀈􀀇􀀈􀀕 􀀖􀀚􀀚􀀂􀀌􀀕􀀇􀀓􀀌􀀂􀀖􀀓􀀎􀀇􀀐􀀈 􀀎􀀔􀀖􀀎 􀀌􀀂q􀀛􀀇􀀌􀀂􀀍 􀀌􀀖p􀀇􀀗 􀀇􀀗􀀂􀀈􀀎􀀇ﬁ􀀓􀀖􀀎􀀇􀀐􀀈 􀀖􀀈􀀗 􀀇􀀈􀀎􀀂􀀌􀀋􀀂􀀈􀀎􀀇􀀐􀀈 P􀀌􀀂􀀗􀀇􀀓􀀎􀀇􀀈􀀕􀀇􀀈􀀗􀀇􀀋􀀇􀀗􀀛􀀖􀀚􀀍 􀀖􀀎 􀀌􀀇􀀍􀀆 􀀌􀀂􀀃􀀖􀀇􀀈􀀍 􀀖 􀀓􀀚􀀇􀀈􀀇􀀓􀀖􀀚 􀀓􀀔􀀖􀀚􀀚􀀂􀀈􀀕􀀂 􀀗􀀛􀀂 􀀎􀀐 􀀇􀀎􀀍 􀀃􀀛􀀚􀀎􀀇􀀑􀀖􀀓􀀎􀀐􀀌􀀇􀀖􀀚 􀀈􀀖􀀎􀀛􀀌􀀂􀀖􀀈􀀗 􀀋􀀖􀀌􀀇􀀖b􀀚􀀂 p􀀌􀀂􀀍􀀂􀀈􀀎􀀖􀀎􀀇􀀐􀀈 􀀌 je􀀖tive􀀞 􀀝􀀐 􀀗􀀂􀀋􀀂􀀚􀀐p 􀀖􀀈􀀗 􀀂􀀋􀀖􀀚􀀛􀀖􀀎􀀂 􀀂xp􀀚􀀖􀀇􀀈􀀖b􀀚􀀂 􀀃􀀖􀀓􀀔􀀇􀀈􀀂 􀀚􀀂􀀖􀀌􀀈􀀇􀀈􀀕 􀀃􀀐􀀗􀀂􀀚􀀍 􀀎􀀔􀀖􀀎 p􀀌􀀂􀀗􀀇􀀓􀀎 􀀎􀀔􀀂 􀀌􀀇􀀍􀀆 􀀐􀀑 􀀖􀀈􀀖p􀀔􀀏􀀚􀀖x􀀇􀀍 􀀛􀀍􀀇􀀈􀀕 􀀌􀀐􀀛􀀎􀀇􀀈􀀂􀀚􀀏 􀀓􀀐􀀚􀀚􀀂􀀓􀀎􀀂􀀗 􀀓􀀚􀀇􀀈􀀇􀀓􀀖􀀚 􀀗􀀖􀀎􀀖 Meth􀀚􀀝s􀀞 W􀀂 􀀖􀀈􀀖􀀚􀀏􀀍􀀂􀀗 􀀖 􀀃􀀖􀀎􀀓􀀔􀀂􀀗 􀀓􀀖􀀍􀀂-􀀓􀀐􀀈􀀎􀀌􀀐􀀚 􀀗􀀖􀀎􀀖􀀍􀀂􀀎 􀀗􀀂􀀌􀀇􀀋􀀂􀀗 􀀑􀀌􀀐􀀃􀀖􀀈􀀐􀀈􀀏􀀃􀀇􀀍􀀂􀀗 􀀂􀀚􀀂􀀓􀀎􀀌􀀐􀀈􀀇􀀓 􀀔􀀂􀀖􀀚􀀎􀀔 􀀌􀀂􀀓􀀐􀀌􀀗􀀍 􀀅􀀑􀀎􀀂􀀌 􀀖pp􀀚􀀏􀀇􀀈􀀕 􀀓􀀔􀀇-􀀍q􀀛􀀖􀀌􀀂􀀗-b􀀖􀀍􀀂􀀗􀀑􀀂􀀖􀀎􀀛􀀌􀀂 􀀍􀀂􀀚􀀂􀀓􀀎􀀇􀀐􀀈􀀉 w􀀂 􀀎􀀌􀀖􀀇􀀈􀀂􀀗 􀀃􀀛􀀚􀀎􀀇p􀀚􀀂 􀀓􀀚􀀖􀀍􀀍􀀇ﬁ􀀓􀀖􀀎􀀇􀀐􀀈 􀀖􀀚􀀕􀀐􀀌􀀇􀀎􀀔􀀃􀀍—􀀇􀀈􀀓􀀚􀀛􀀗􀀇􀀈􀀕􀀚􀀐􀀕􀀇􀀍􀀎􀀇􀀓 􀀌􀀂􀀕􀀌􀀂􀀍􀀍􀀇􀀐􀀈􀀉 􀀗􀀂􀀓􀀇􀀍􀀇􀀐􀀈 􀀎􀀌􀀂􀀂􀀍􀀉 􀀌􀀖􀀈􀀗􀀐􀀃 􀀑􀀐􀀌􀀂􀀍􀀎􀀍􀀉 XGB􀀐􀀐􀀍􀀎􀀉 􀀖􀀈􀀗 􀀖 􀀍􀀎􀀖􀀓􀀆􀀇􀀈􀀕􀀂􀀈􀀍􀀂􀀃b􀀚􀀂 􀀒􀀐􀀗􀀂􀀚 p􀀂􀀌􀀑􀀐􀀌􀀃􀀖􀀈􀀓􀀂 w􀀖􀀍 􀀂􀀋􀀖􀀚􀀛􀀖􀀎􀀂􀀗 􀀛􀀍􀀇􀀈􀀕 􀀅􀀊􀀁􀀉 􀀍􀀂􀀈􀀍􀀇􀀎􀀇􀀋􀀇􀀎􀀏􀀉 􀀍p􀀂􀀓􀀇ﬁ􀀓􀀇􀀎􀀏􀀉 p􀀌􀀂􀀓􀀇􀀍􀀇􀀐􀀈􀀉 􀀖􀀈􀀗 F1-􀀍􀀓􀀐􀀌􀀂 􀀘H􀀅P 􀀋􀀖􀀚􀀛􀀂􀀍 w􀀂􀀌􀀂 􀀛􀀍􀀂􀀗 􀀎􀀐 􀀖􀀍􀀍􀀂􀀍􀀍 􀀃􀀐􀀗􀀂􀀚 􀀂xp􀀚􀀖􀀇􀀈􀀖b􀀇􀀚􀀇􀀎􀀏 􀀈es􀀛lts􀀞 􀀝􀀔􀀂 b􀀂􀀍􀀎-p􀀂􀀌􀀑􀀐􀀌􀀃􀀇􀀈􀀕 􀀃􀀐􀀗􀀂􀀚 􀀖􀀓􀀔􀀇􀀂􀀋􀀂􀀗 􀀖􀀈 􀀅􀀊􀀁 􀀐􀀑 0 79􀀉 􀀗􀀂􀀃􀀐􀀈􀀍􀀎􀀌􀀖􀀎􀀇􀀈􀀕􀀔􀀇􀀕􀀔 􀀗􀀇􀀍􀀓􀀌􀀇􀀃􀀇􀀈􀀖􀀎􀀇􀀐􀀈 􀀖􀀈􀀗 b􀀖􀀚􀀖􀀈􀀓􀀂􀀗 􀀍􀀂􀀈􀀍􀀇􀀎􀀇􀀋􀀇􀀎􀀏/􀀍p􀀂􀀓􀀇ﬁ􀀓􀀇􀀎􀀏 K􀀂􀀏 p􀀌􀀂􀀗􀀇􀀓􀀎􀀐􀀌􀀍􀀇􀀈􀀓􀀚􀀛􀀗􀀂􀀗 􀀔􀀂􀀖􀀚􀀎􀀔􀀓􀀖􀀌􀀂 􀀛􀀎􀀇􀀚􀀇􀀍􀀖􀀎􀀇􀀐􀀈 p􀀖􀀎􀀎􀀂􀀌􀀈􀀍􀀉 􀀖􀀕􀀂􀀉 􀀍􀀐􀀓􀀇􀀐􀀂􀀓􀀐􀀈􀀐􀀃􀀇􀀓 p􀀌􀀐x􀀏(􀀓􀀐p􀀖􀀏􀀃􀀂􀀈􀀎 􀀚􀀂􀀋􀀂􀀚)􀀉 􀀖􀀈􀀗 􀀍p􀀂􀀓􀀇ﬁ􀀓 􀀗􀀇􀀖􀀕􀀈􀀐􀀍􀀎􀀇􀀓 􀀓􀀐􀀗􀀂􀀍 􀀌􀀂􀀚􀀖􀀎􀀂􀀗 􀀎􀀐 􀀖􀀚􀀚􀀂􀀌􀀕􀀇􀀓 􀀓􀀐􀀈􀀗􀀇􀀎􀀇􀀐􀀈􀀍􀀋􀀚􀀜􀀖l􀀛si􀀚􀀜􀀞 􀀝􀀔􀀇􀀍 􀀍􀀎􀀛􀀗􀀏 􀀗􀀂􀀃􀀐􀀈􀀍􀀎􀀌􀀖􀀎􀀂􀀍 􀀎􀀔􀀂 p􀀐􀀎􀀂􀀈􀀎􀀇􀀖􀀚 􀀐􀀑 􀀇􀀈􀀎􀀂􀀌p􀀌􀀂􀀎􀀖b􀀚􀀂 􀀃􀀖􀀓􀀔􀀇􀀈􀀂􀀚􀀂􀀖􀀌􀀈􀀇􀀈􀀕 􀀖pp􀀌􀀐􀀖􀀓􀀔􀀂􀀍 􀀎􀀐 􀀍􀀛pp􀀐􀀌􀀎 􀀎􀀔􀀂 􀀂􀀖􀀌􀀚􀀏 􀀇􀀗􀀂􀀈􀀎􀀇ﬁ􀀓􀀖􀀎􀀇􀀐􀀈 􀀐􀀑 􀀇􀀈􀀗􀀇􀀋􀀇􀀗􀀛􀀖􀀚􀀍 􀀖􀀎 􀀔􀀇􀀕􀀔􀀌􀀇􀀍􀀆 􀀐􀀑 􀀖􀀈􀀖p􀀔􀀏􀀚􀀖x􀀇􀀍 􀀝􀀔􀀂􀀍􀀂 􀀎􀀐􀀐􀀚􀀍 􀀓􀀖􀀈 􀀂􀀈􀀔􀀖􀀈􀀓􀀂 􀀓􀀚􀀇􀀈􀀇􀀓􀀖􀀚 􀀌􀀇􀀍􀀆 􀀍􀀎􀀌􀀖􀀎􀀇ﬁ􀀓􀀖􀀎􀀇􀀐􀀈 􀀖􀀈􀀗􀀇􀀈􀀑􀀐􀀌􀀃 p􀀌􀀂􀀋􀀂􀀈􀀎􀀇􀀋􀀂 􀀍􀀎􀀌􀀖􀀎􀀂􀀕􀀇􀀂􀀍 􀀇􀀈 􀀌􀀐􀀛􀀎􀀇􀀈􀀂 p􀀌􀀖􀀓􀀎􀀇􀀓􀀂 K􀀁􀀇􀀊􀀌􀀈􀀂􀀍 􀀕􀀜􀀕phyl􀀕xis p􀀙e􀀝i􀀖ti􀀚􀀜, st􀀕􀀖􀀗e􀀝 m􀀕􀀖hi􀀜e le􀀕􀀙􀀜i􀀜􀀘 m􀀚􀀝el, 􀀕􀀝mi􀀜ist􀀙􀀕tive he􀀕lth􀀖􀀕􀀙e 􀀝􀀕t􀀕, el􀀝e􀀙ly p􀀚p􀀛l􀀕ti􀀚􀀜, he􀀕lth􀀖􀀕􀀙e 􀀛tilis􀀕ti􀀚􀀜 p􀀕tte􀀙􀀜s, 􀀕􀀙tiﬁ􀀖i􀀕l i􀀜telli􀀘e􀀜􀀖e, 􀀕lle􀀙􀀘y 􀀙is􀀗 st􀀙􀀕tiﬁ􀀖􀀕ti􀀚􀀜 I􀀈􀀇􀀂􀀗􀀄􀀕􀀆􀀇􀀅􀀗􀀈 􀀁􀀂􀀃􀀄􀀅􀀆􀀇􀀃􀀈􀀉􀀊 􀀉􀀊 􀀃 􀀊􀀌􀀍􀀌􀀎􀀌 􀀃􀀂􀀏 􀀄􀀐􀀑􀀌􀀂􀀑􀀉􀀃􀀇􀀇􀀆 􀀇􀀉􀀒􀀌􀀓􀀑􀀅􀀎􀀌􀀃􀀑􀀌􀀂􀀉􀀂􀀔 􀀃􀀇􀀇􀀌􀀎􀀔􀀉􀀕 􀀎􀀌􀀃􀀕􀀑􀀉􀀐􀀂 􀀑􀀅􀀃􀀑 􀀕􀀃􀀂􀀐􀀕􀀕􀀖􀀎 􀀎􀀃􀀄􀀉􀀏􀀇􀀆 􀀃􀀂􀀏 􀀏􀀌􀀗􀀃􀀂􀀏􀀊 􀀉􀀗􀀗􀀌􀀏􀀉􀀃􀀑􀀌 􀀗􀀌􀀏􀀉􀀕􀀃􀀇 􀀉􀀂􀀑􀀌􀀎􀀍􀀌􀀂􀀑􀀉􀀐􀀂􀀘 􀀙􀀌􀀊􀀄􀀉􀀑􀀌 􀀉􀀑􀀊 􀀕􀀇􀀉􀀂􀀉􀀕􀀃􀀇􀀖􀀎􀀔􀀌􀀂􀀕􀀆􀀚 􀀄􀀎􀀌􀀏􀀉􀀕􀀑􀀉􀀂􀀔 􀀑􀀅􀀌 􀀐􀀂􀀊􀀌􀀑 􀀐􀀒 􀀛􀀎􀀊􀀑􀀓􀀑􀀉􀀗􀀌 􀀃􀀂􀀃􀀄􀀅􀀆􀀇􀀃􀀈􀀉􀀊 􀀜􀀑􀀅􀀌 􀀛􀀎􀀊􀀑 􀀎􀀌􀀕􀀐􀀎􀀏􀀌􀀏 􀀏􀀉􀀃􀀔􀀂􀀐􀀊􀀉􀀊 􀀐􀀒􀀃􀀂􀀃􀀄􀀅􀀆􀀇􀀃􀀈􀀉􀀊􀀝 􀀎􀀌􀀗􀀃􀀉􀀂􀀊 􀀃 􀀊􀀉􀀔􀀂􀀉􀀛􀀕􀀃􀀂􀀑 􀀕􀀅􀀃􀀇􀀇􀀌􀀂􀀔􀀌􀀘 􀀞􀀅􀀉􀀊 􀀏􀀉􀀒􀀛􀀕􀀖􀀇􀀑􀀆 􀀃􀀎􀀉􀀊􀀌􀀊 􀀒􀀎􀀐􀀗 􀀑􀀅􀀌 􀀕􀀐􀀗􀀄􀀇􀀌􀀈􀀉􀀂􀀑􀀌􀀎􀀄􀀇􀀃􀀆 􀀐􀀒 􀀉􀀂􀀏􀀉􀀍􀀉􀀏􀀖􀀃􀀇 􀀎􀀉􀀊 􀀒􀀃􀀕􀀑􀀐􀀎􀀊􀀚 􀀉􀀂􀀕􀀇􀀖􀀏􀀉􀀂􀀔 􀀖􀀂􀀏􀀌􀀎􀀇􀀆􀀉􀀂􀀔 􀀅􀀌􀀃􀀇􀀑􀀅 􀀕􀀐􀀂􀀏􀀉􀀑􀀉􀀐􀀂􀀊􀀚􀀄􀀐􀀇􀀆􀀄􀀅􀀃􀀎􀀗􀀃􀀕􀀆􀀚 􀀃􀀂􀀏 􀀊􀀐􀀕􀀉􀀐􀀌􀀕􀀐􀀂􀀐􀀗􀀉􀀕 􀀏􀀌􀀑􀀌􀀎􀀗􀀉􀀂􀀃􀀂􀀑􀀊􀀚 w􀀅􀀉􀀕􀀅 􀀃􀀎􀀌 􀀐􀀒􀀑􀀌􀀂 􀀂􀀐􀀑 􀀃􀀏􀀌q􀀖􀀃􀀑􀀌􀀇􀀆􀀕􀀃􀀄􀀑􀀖􀀎􀀌􀀏 􀀉􀀂 􀀕􀀖􀀎􀀎􀀌􀀂􀀑 􀀊􀀕􀀎􀀌􀀌􀀂􀀉􀀂􀀔 􀀄􀀎􀀐􀀑􀀐􀀕􀀐􀀇􀀊 􀀜1􀀚 2􀀝􀀘 I􀀂 􀀗􀀉􀀏􀀏􀀇􀀌􀀓􀀃􀀔􀀌􀀏 􀀃􀀂􀀏 􀀌􀀇􀀏􀀌􀀎􀀇􀀆 􀀄􀀐􀀄􀀖􀀇􀀃􀀑􀀉􀀐􀀂􀀊􀀚 􀀃􀀂􀀃􀀄􀀅􀀆􀀇􀀃􀀈􀀉􀀊 􀀉􀀊 􀀐􀀒􀀑􀀌􀀂 􀀗􀀐􀀎􀀌 􀀊􀀌􀀍􀀌􀀎􀀌 􀀃􀀂􀀏􀀃􀀊􀀊􀀐􀀕􀀉􀀃􀀑􀀌􀀏 w􀀉􀀑􀀅 􀀅􀀉􀀔􀀅􀀌􀀎 􀀅􀀐􀀊􀀄􀀉􀀑􀀃􀀇􀀉􀀊􀀃􀀑􀀉􀀐􀀂 􀀎􀀃􀀑􀀌􀀊 􀀕􀀐􀀗􀀄􀀃􀀎􀀌􀀏 􀀑􀀐 􀀆􀀐􀀖􀀂􀀔􀀌􀀎 􀀔􀀎􀀐􀀖􀀄􀀊􀀘 I􀀂􀀊􀀌􀀕􀀑 F􀀂􀀗􀀈􀀇􀀅􀀃􀀂􀀌 􀀅􀀈 A􀀓􀀓􀀃􀀂􀀉􀀒 0􀀚 f􀀂􀀗􀀈􀀇􀀅􀀃􀀂􀀌􀀅􀀈 .􀀗􀀂􀀉 􀀍􀀌􀀂􀀐􀀗􀀊 􀀃􀀂􀀏 􀀗􀀌􀀏􀀉􀀕􀀃􀀑􀀉􀀐􀀂􀀊􀀚 􀀄􀀃􀀎􀀑􀀉􀀕􀀖􀀇􀀃􀀎􀀇􀀆 􀀃􀀂􀀃􀀇􀀔􀀌􀀊􀀉􀀕􀀊 􀀃􀀂􀀏 􀀃􀀂􀀑􀀉

FAQs

What is the document’s main goal?

To predict first-time anaphylaxis in elderly people using a stacked machine learning method trained on population registry data.

What type of data is used to build the prediction models?

Information drawn from population registries, linked to individuals to create model features for risk estimation.

How is the model evaluated in the study?

Performance is assessed using standard predictive evaluation metrics and related analyses, including inspection of which factors contribute most to prediction.

Aug 4, 2026

Upload by Theodora