Table 2.

Reliable Diagnosis Challenges taxonomy (total count)

Category	GPT	Human	P value
Challenging disease presentation
Atypical presentation	513 (94.1%)	211 (38.7%)	<0.001
Non-specific symptoms and signs	82 (15.0%)	69 (12.7%)	0.25
Unfamiliar/outside specialty	161 (29.5%)	61 (11.2%)	<0.001
Findings masking/mimicking another diagnosis	383 (70.3%)	210 (38.5%)	<0.001
Red herring misleading findings	83 (15.2%)	91 (16.7%)	0.51
Rapidly progressive course	6 (1.1%)	21 (3.9%)	0.003
Slowly evolving blunting onset perception	54 (9.9%)	39 (7.2%)	0.10
Deceptively benign course	24 (4.4%)	8 (1.5%)	0.004
Patient factors
Language/communication barriers	9 (1.7%)	14 (2.6%)	0.29
Signal: noise - patients with multiple other symptoms or diagnoses	24 (4.4%)	23 (4.2%)	0.88
Failure to share data (to be forthcoming with symptoms or their severity)	79 (14.5%)	10 (1.8%)	<0.001
Failure to follow-up	64 (11.7%)	4 (0.7%)	<0.001
Testing challenges
Test not available due to geography, access, cost	58 (10.6%)	5 (0.9%)	<0.001
Logistical issues in scheduling, performing	31 (5.7%)	2 (0.4%)	<0.001
False positive/negative test limitations	54 (9.9%)	38 (7.0%)	0.08
Performance/interpretation failures	444 (81.5%)	145 (26.6%)	<0.001
Equivocal results/interpretation	99 (18.2%)	44 (8.1%)	<0.001
Test follow-up issues (eg, tracking pending results)	39 (7.2%)	5 (0.9%)	<0.001
Stressors
Time constraints for clinicians and patients	32 (5.9%)	2 (0.4%)	<0.001
Discontinuities of care	137 (25.1%)	1 (0.2%)	<0.001
Fragmentation of care	36 (6.6%)	4 (0.7%)	<0.001
Memory reliance/challenges	20 (3.7%)	3 (0.6%)	<0.001
Broader challenges
Recognition of acuity/severity	91 (16.7%)	43 (7.9%)	<0.001
Diagnosis of complications	16 (2.9%)	35 (6.4%)	0.01
Recognition of failure to respond to therapy	23 (4.2%)	28 (5.1%)	0.47
Diagnosis of the underlying aetiological cause	44 (8.1%)	149 (27.3%)	<0.001
Recognising misdiagnosis occurrence	77 (14.1%)	85 (15.6%)	0.50
Unclear	0 (0.0%)	3 (0.6%)	0.08