Introduction
Diagnostic errors and clinical decision support system
-
A delayed diagnosis
-
A wrong diagnosis
-
A missed diagnosis [4]
Rare diseases, difficult-to-diagnose cases, and clinical diagnosis support systems
Country | Prevalence | Source |
---|---|---|
The EU | < 1 person in 2000 | EU research on rare diseases |
Japan | Not defined | Act on Medical Care for Patients with Intractable Diseases |
The UK | < 1 person in 2000 | The UK Rare Diseases Framework |
The US | < 50,000 persons in the US | Rare Diseases Act of 2002 |
Main objectives of the clinical decision support system
-
To support differential diagnoses performed by internists and general practitioners.
-
To prevent diagnostic errors made by physicians
Main features of the clinical decision support system
A physician inputs a patient’s symptoms, findings, and test results to the system, and the system outputs a ranking list of possible diseases.
-
Subjective symptoms
-
Objective findings
-
Physical findings
-
Laboratory test results
-
Imaging test results
-
Other Information
-
A ranking list of possible diseases
Example of the clinical decision support system
Inputted symptoms | Score | Predicted diseases | |||
---|---|---|---|---|---|
a | Fever | 1 | 1.61 | Acute HIV-1 infection | |
b | Headache | 2 | 1.51 | Polyneuropathy | |
c | Sore throat | 3 | 0.91 | Acute viral meningitis | |
d | Consciousness indistinctness | 4 | 0.88 | West Nile fever | |
e | Chills | 5 | 0.77 | Cat-scratch disease | |
f | Muscles ache | 6 | 0.46 | Acute Q fever | |
g | Swallowing pain | → | 7 | 0.23 | Hepatitis A |
h | Pharyngolaryngeal abnormality | 8 | 0.21 | Chronic fatigue syndrome | |
i | Aphasia | 9 | 0.13 | Sepsis | |
j | Apraxia | 10 | 0.12 | Toxoplasmosis | |
k | Fatigue | … | |||
l | Muscle weakness | ||||
m | Anorexia | ||||
n | Weight loss | ||||
o | Dementia |
Background
Differential diagnosis process by physicians and learning-to-rank by machines
-
Pointwise approach
-
Pairwise approach
-
Listwise approach [10]
-
Pointwise approach:
-
Score one differential disease at a time.
-
-
Pairwise approach:
-
Compare two differential diseases at a time.
-
-
Listwise approach:(1)Recall multiple differential diseases(2)Refine the recalled differential diseases(3)Rank the refined differential diseases
Case data for clinical decision support system
-
Medical textbooks
-
Medical treatises
-
Medical articles
-
Case reports
-
Symptoms
-
Confirmed disease(s)
-
Differential diseases (related or to be excluded)
Rare diseases and difficult-to-diagnose cases that internists and general practitioners may close encounter in actual cases.
Information retrieval and clinical decision support system
Items | Information retrieval (Ex: Google scholar) | Clinical decision support system |
---|---|---|
Objectives | Get medical literature for target diseases | Get possible diseases |
Target data | Medical literature | ← |
Method of retrieving target data | Web crawlers, etc | Selection by physicians |
Framework | Learning-to-rank | ← |
Input data | Symptoms, Diseases | Symptoms |
Output data | Ranking list of useful medical literatures | Ranking list of possible diseases |
Evaluation method | Subjective evaluation | Objective evaluation |
Physicians | Case reports | |
Evaluation Functions |
Conventional clinical decision support systems
-
The predictive algorithms are LTR with a pointwise approach.
-
These algorithms are less affinitive to the DDx process by experienced physicians.
-
The case data does not include information on differential diseases.
-
These algorithms do not use the relationship between confirmed disease(s) and differential diseases.
Figures and tables
Design
Design principles
-
The prediction algorithms should be higher affinitive to the DDx process by experienced physicians.
-
The case data should include not only information on confirmed disease(s) but also information on differential diseases.
-
These algorithms should utilize the relationship between confirmed disease(s) and differential diseases.
-
To focus on commonalities between IR and CDSS, utilize various IR technologies for CDSS.
Library for learning-to-rank
Case date for learning-to-rank with the listwise approach
-
Symptoms
-
Confirmed disease
Code | Observed symptoms | Code | Diseases | ||
---|---|---|---|---|---|
a | Fever | Fever | 548 | Acute HIV-1 infection | |
b | Head | Headache | |||
c | Sore | Sore throat | |||
d | Myalg | Muscles ache | |||
e | Fatig | Fatigue | → | ||
f | Weigh | Weight loss | |||
g | Arthralg | Arthralgia | |||
h | Diarrh | Diarrhea | |||
i | Lymphn | Lymphadenopathy | |||
j | Mening | Meningitis | |||
… |
-
Symptoms
-
Confirmed disease(s) and these scores
-
Differential diseases (related or to be excluded) and these scores
Code | Observed symptoms | Scores | Code | Diseases | ||
---|---|---|---|---|---|---|
a | Fever | Fever | 17.078 | 548 | Acute HIV-1 infection | |
b | Head | Headache | 12.086 | 296 | Acute hepatitis | |
c | Sore | Sore throat | 11.250 | 102 | Toxoplasmosis | |
d | Myalg | Muscles ache | 11.000 | 491 | Severe fever with thrombocytopenia syndrome (SFTS) | |
e | Fatig | Fatigue | → | 10.836 | 391 | Osteomyelitis |
f | Weigh | Weight loss | 10.836 | 589 | Polyneuropathy | |
g | Arthralg | Arthralgia | 10.836 | 641 | Coccidioidomycosis | |
h | Diarrh | Diarrhea | 10.664 | 627 | Cat-scratch disease | |
i | Lymphn | Lymphadenopathy | 10.500 | 541 | Infectious endocarditis | |
j | Mening | Meningitis | 10.414 | 989 | Dengue (hemorrhagic) fever | |
… | … |
Evaluation
Evaluation purposes
-
The Machine Learning (ML) performance
-
The ML performance of the system is superior to the conventional system.
-
-
The Differential Diagnostic (DDx) performance
-
The DDx performance of the system is superior to the conventional system.
-
The DDx performance of the system is useful to support the DDx process by physicians.
-
The Clinical Decision Support system (CDSS) is useful in preventing diagnostic errors by physicians.
-
-
Loss function:UPPER CASE (ex: NDCG, MSE, etc.)
-
Evaluation function: lower case (ex: ndcg, mse, etc.)
The compared system
-
The main objective is to propose the prediction algorithm (Learning-to-Rank; listwise approach) for CDSS. In the interest of fairness, the comparison conditions (training data, etc.), except for the algorithm, must be the same. However, these systems' algorithms and training data are not publicly available.
-
Each CDSS has different objectives and target diseases.
Evaluation criteria for differential diagnostic performance
-
Physicians decide the final confirmed disease(s) by themselves, using the predicted diseases of CDSS as a reference.
Case selection criteria for evaluation of differential diagnostic performance
-
Our main target diseases are rare diseases and difficult-to-diagnose cases that internists and general practitioners may close encounter in clinical practice. However, the probability of encountering these diseases is low.
-
For correct evaluation, it is important to evaluate with validated cases.
Evaluation: machine learning performance
Evaluation method
-
Learning curves
-
Value of evaluation function
Evaluation results and discussion
Loss functions | Evaluation functions | |||
---|---|---|---|---|
ndcg | ndcg@5 | ndcg@10 | ndcg@20 | |
A-NDCG | 0.7098 | 0.6205 | 0.6485 | 0.6680 |
MSE | 0.5835 | 0.4470 | 0.4845 | 0.5139 |
-
The number of epochs in training was larger for MSE.
-
However, the training time was longer for A-NDCG.
-
The memory space requirement was larger for A-NDCG.
-
We found that the prediction model with A-NDCG tended to overfit.
-
For LTR, we found that mse was not a suitable evaluation function.
-
The value of the evaluation functions was consistently higher for A-NDCG.
-
Hyperparameters tuning with Bayesian optimization
-
Change of the neural network configuration
-
Number of layers
-
Activation function
-
Optimizer algorithm
-
Evaluation: differential diagnosis performance
Evaluation method
-
Predicted diseases
-
Inputted symptoms and predicted diseases
-
Inputted symptoms and the target disease's ranking
Evaluation results and discussion
Disease with characteristic symptoms
-
Acute intermittent porphyria (AIP)
A-NDCG | MSE | |
---|---|---|
1 | Acute intermittent porphyria | Acute intermittent porphyria |
2 | Diabetic coma imminent state | Enterohemorrhagic e. coli (EHEC) infection |
3 | Pesticide poisoning, Organophosphate toxicity | Visceral rupture |
4 | Lead poisoning (almost chronic) | Fibromyalgia (fibrositis) |
5 | Heat stroke (hyperthermia) | Cancerous peritonitis |
6 | Cytomegalovirus infection | Withdrawal symptoms of alcohol and drugs |
7 | Visceral rupture | Colorectal cancer |
8 | Hyponatremia | Irritable bowel syndrome, Functional dyspepsia (FD) |
9 | Portal vein obstruction | Drugs (laxatives, etc.) |
10 | Acetaminophen poisoning | Eating disorder |
… |
Difficult-to-diagnose case with few characteristic symptoms
-
Acute HIV-1 infection
-
Acute viral meningitis
A-NDCG | MSE | |
---|---|---|
1 | Acute HIV-1 infection | Epidemic hepatitis A |
2 | Polyneuropathy | Acute Q fever |
3 | Acute viral meningitis | Acute pharyngitis |
4 | West Nile fever | Polyneuropathy |
5 | Cat-scratch disease | Lymphocytic choriomeningitis |
6 | Acute Q fever | Herpes labialis |
7 | Epidemic hepatitis A | Side effects of interferon |
8 | Chronic fatigue syndrome | Sepsis |
9 | Sepsis | Chronic fatigue syndrome |
10 | Toxoplasmosis | Retropharyngeal infection |
… |
Case with diagnostic errors
-
Subacute bacterial endocarditis (SBE)
-
Acute bacterial endocarditis
-
Infectious endocarditis
-
Mixed cryoglobulinemia
Predicted diseases | Classification | |
---|---|---|
1 | Zieve syndrome | |
2 | Disseminated intravascular coagulation | |
3 | Chronic hepatitis | |
4 | Wilson's disease | |
5 | Acute hepatitis | |
6 | Hepatic amyloidosis | |
7 | Infectious endocarditis | Related disease |
8 | (Compensated/uncompensated) liver cirrhosis | |
9 | Subacute bacterial endocarditis | Related disease |
10 | Gastric cancer | |
… |
Predicted diseases | Classification | |
---|---|---|
1 | Mixed cryoglobulinemia | Misdiagnosed disease |
2 | Chronic hepatitis | |
3 | Subacute bacterial endocarditis | Related disease |
4 | Hepatic amyloidosis | |
5 | Rapidly progressive glomerulonephritis syndrome | |
6 | Acute bacterial endocarditis | Related disease |
7 | Infectious endocarditis | Related disease |
8 | Polyarteritis nodosa | |
9 | Autoimmune hemolytic anemia | |
10 | Disseminated intravascular coagulation | |
… |
Conclusion
Evaluation results
Differential diagnosis process by physicians and learning to rank by machines
Case data and information retrieval
Case data (= training data) | Predicted results | |
---|---|---|
X: explanatory variables | Observed symptoms | Inputted symptoms |
y: explained variables | Confirmed disease(s) and those score(s) & Differential diseases (related or to be excluded) and those scores | Predicted diseases and those scores |
-
Technology for predicting diseases, such as Learning-to-Rank (LTR)
-
Technology for text-mining information on diseases from literatures
-
Technology for converting text-mining data to the symptoms and diseases
Potentials for clinical decision support system
-
Recall rare diseases
-
Support differential diagnoses for difficult-to-diagnose cases
-
Prevent diagnostic errors
Evolution into explainable clinical decision support system
-
The affinity between Differential Diagnosis (DDx) processes by experienced physicians and LTR with the listwise approach
-
The similarity between case data (= training data) and predicted results
-
The simple neural network
-
The number of internal hiding layers is one.
-
The number of learnable times (epochs) is relatively small.
-