Back

Can Large Language Models Diagnose Primary Immunodeficiency from Patient-Described Symptoms?

Reteig, L. C.; Woloshin, S.; Maglione, P. J.; Farmer, J. R.; Ong, M.-S.

2026-05-27 allergy and immunology
10.64898/2026.05.26.26353818 medRxiv
Show abstract

Patients with primary immunodeficiency (PID) often face prolonged diagnostic delays and may increasingly turn to large language models (LLMs) to interpret their symptoms during this period. We evaluated whether an LLM could recognize PID from symptom descriptions derived from interviews with 21 PID patients. In a prior study, we showed that GPT-4o identified PID in 96% of cases when prompted with physician-written patient histories (Rider et al., JACI, 2024). Here, when prompted with symptom descriptions in patients' own words, GPT-5 identified PID in only 7 cases (33%), although it more broadly suggested immune system issues in 18 cases (81%). The gap between these findings indicates that LLMs are sensitive to the language and framing of symptom descriptions, performing substantially worse when patients describe their own symptoms in everyday language than when clinicians summarize patient histories in structured medical terms. This study underscores the need to carefully evaluate how LLMs are used in patient-facing applications.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

1
International Journal of Medical Informatics
25 papers in training set
Top 0.1%
12.6%
2
Journal of The Royal Society Interface
189 papers in training set
Top 0.4%
7.0%
3
PLOS Digital Health
91 papers in training set
Top 0.3%
6.5%
4
Scientific Reports
3102 papers in training set
Top 16%
6.5%
5
PLOS ONE
4510 papers in training set
Top 26%
6.5%
6
Cell Reports Methods
141 papers in training set
Top 0.4%
5.0%
7
Proceedings of the National Academy of Sciences
2130 papers in training set
Top 13%
5.0%
8
eLife
5422 papers in training set
Top 24%
3.7%
50% of probability mass above
9
Bioinformatics Advances
184 papers in training set
Top 1%
3.7%
10
mSphere
281 papers in training set
Top 2%
3.1%
11
npj Digital Medicine
97 papers in training set
Top 2%
2.8%
12
Clinical Infectious Diseases
231 papers in training set
Top 2%
2.8%
13
iScience
1063 papers in training set
Top 11%
1.9%
14
Journal of the American Medical Informatics Association
61 papers in training set
Top 1%
1.7%
15
Frontiers in Digital Health
20 papers in training set
Top 0.9%
1.3%
16
Life Science Alliance
263 papers in training set
Top 0.6%
1.3%
17
European Respiratory Journal
54 papers in training set
Top 1%
1.0%
18
The American Journal of Human Genetics
206 papers in training set
Top 3%
1.0%
19
Journal of Medical Internet Research
85 papers in training set
Top 4%
1.0%
20
Cell Reports Medicine
140 papers in training set
Top 6%
0.9%
21
Nature Medicine
117 papers in training set
Top 5%
0.8%
22
Genetics
225 papers in training set
Top 4%
0.8%
23
Computers in Biology and Medicine
120 papers in training set
Top 5%
0.7%
24
Epidemics
104 papers in training set
Top 2%
0.7%
25
Science Translational Medicine
111 papers in training set
Top 7%
0.7%
26
European Journal of Human Genetics
49 papers in training set
Top 1%
0.7%
27
Journal of Clinical Microbiology
120 papers in training set
Top 2%
0.5%
28
PLOS Computational Biology
1633 papers in training set
Top 28%
0.5%
29
Healthcare
16 papers in training set
Top 2%
0.5%
30
Communications Biology
886 papers in training set
Top 31%
0.5%