# 醫學資訊概論3.Biomedical Decision Making-Probabilistic Clinical Reasoning

[簡報在這裡](https://elearn2.fju.edu.tw/course/234105/learning-activity#/1242096)
## 小考題目及解答:
- 選擇題
1. Which is not a nature(characteristic,特性)of clinical data?
- [ ] Uncertain
- [ ] Imprecision
- [X] certain and precision
- [ ] Fuzziness
2. Which is to measure the likelihood that a diseased patient has a negative test result?
- [X] false-nagative(FNR)
- [ ] prevalence
- [ ] true-nagative rate(TNR)
- [ ] true-positive rate(TPR)
3. Which is more correct statement about 'probability' referring to?
- [ ] degress of incompleteness
- [X] degrees of certainty
- [ ] degrees of consistency
4. Which is correct statement about 'prior probability' or called pretest probability?
- [X] The probability that a clinician makes an initial judgement(hypotheses) only based on what he or she knows from talking with the patient.
- [ ] The probability that a clinician makes an initial judement(hypotheses) after he or she performs a diagnostic test such as a blood test.
- [ ] The probability that a clinician makes a refine judgement(hypotheses) only based on what he or she knows.
5. Which is correct statement about 'post-test probability' or called 'posterior probability'?
- [ ] The probability that a clinician makes an initial judgement(hypotheses) only based on what he or she knows from talking with the patient.
- [ ] The probability that a clinician makes an initial judement(hypotheses) after he or she performs a diagnostic test such as a blood test.
- [X] The probability that a clinician makes a refine judgement(hypotheses) after he or she performs a diagnostic as a blood test.
6. 疾病檢測(Test)結果，特異性(specificity)指的是下列哪一種？
- [ ] true-positive
- [X] true-negative rate
- [ ] flase-positive
- [ ] flase-negative rate
7. 疾病檢測結果，敏感性(sensitivity)指的是下列哪一種？
- [X] true-positive
- [ ] true-negative rate
- [ ] flase-positive
- [ ] flase-negative rate
8. 疾病檢測結果，流行(盛行)率(prevalence rate)的計算式是下列哪一種？
- [ ] True negative/(Total test populstion)
- [ ] True positive/(Total test population)
- [ ] Total pretest negative/(Total test population)
- [X] Total pretest positive/(Total test population)
9. Which step of decision analysis means "to test the conclusions of the analysis"?
- [X] Use sensitivity analysis
- [ ] Calculate the expected value of each decision alternative
- [ ] Choose the decision alternative with the highest expected value
- [ ] Create a decision tree
10. Which is an incorrect statement about hypothetico-deductive(假設-推論) approach(HDA) to treat patient's illness(治療病人的疾病)？
- [ ] It is an iteration process and sequential, staged data collection, followed by data interpretation and the generation of hypotheses.
- [X] The goal often can not narrow/generate a single diagnosis which serves as the basis for providing a more certainty of treatment to the patient.
- [ ] It is a process of medical data collection and treatment id intimately tied to an ongoing process of hypothesis generation and refinement.
11. Which is a method commonly used for representing and comparing the expected outcomes(期望結果)of each decision alternative
- [ ] The threshold probility
- [x] Decision tree
- [ ] Decision forest
- [ ] Bayes'theorem
- 計算題
1. The polymerase chain reaction (聚合酶酵素連鎖反應)test, 簡稱PCR test 為一種基因改造生物檢測法, is a gene-amplication technique that can diagnose HIV test 目前改良後作為COVID-19的檢測。

| PCR test |COVID-19檢測<br>為正常總人數|COVID-19檢測<br>為不正常總人數|
| --------- | -------- | -------- |
|檢測結果為正常| 160 TN | 20 FP |
|檢測結果為異常| 40 FN | 80 TP |

根據上表檢測結果，分別計算下列三項檢測效能指標(計算至小數一位)。
A) True-positive rate(TPR) =66.6%
- $TPR= \frac{80}{80+40}=\frac{2}{3}$

B) Ture-negative rate(TNR) =88.8%
- $TNR= \frac{160}{160+20}=\frac{8}{9}$

C) Prevalence檢測有病除以總數 =33%
- $流行率= \frac{80+20}{160+40+20+80}=\frac{1}{3}$

他們及其他檢測效能指標(Performance Indications) 之對應公式明列如下:
- $TPR= \frac{TP}{TP+FN}$
- $TNR= \frac{TN}{TN+FP}$
- $prevalence(流行率)= \frac{TP+FP}{TP+FN+TN+FP}$
- $PV+= \frac{TP}{TP+FP}$
- $PV-= \frac{TN}{TN+TP}$
- $Correctness(正確性)= \frac{TP+TN}{TP+TN+FN+FP}$

2. 根據以下決策樹(Decision Tree) ，計算Surgery之期望值並指出是否應進行Surgery?

digraph summary {
    edge [style="" dir="none"]
    st [label="start" fontsize=30 shape=box]
    op1 [label="D" shape=octagon]
    op2 [label="C" shape=octagon]
    op3 [label="B" shape=octagon]
    op4 [label="A" shape=octagon]
    re1 [label="Poor mobility
    6Years" shape=box]
    re2 [label="Full mobility
    10Years" shape=box]
    re3 [label="Whellchair-bound
    3Years" shape=box]
    re4 [label="Death
    0Years" shape=box]
    
    st->op1[label="Surgery" fontsize=20]
    st->re1[label="No surgery" fontsize=20]
    op1->op2[label="Survival
p=0.95" fontsize=10]
    op1->re4[label="Operative death
p=0.05" fontsize=10]
    op2->op3[label="No infection
p=0.95" fontsize=10]
    op2->op4[label="Infection
p=0.05" fontsize=10]
    op3->re1[label="p=0.4" fontsize=10]
    op3->re2[label="p=0.6" fontsize=10]
    op4->re3[label="p=0.95" fontsize=10]
    op4->re4[label="p=0.05" fontsize=10]
    }

A = 0.05x0+0.95x3=2.85
B = 10x0.6+6x0.4=8.4
C = Ax0.05+Bx0.95=8.1225
D = Cx0.95+0x0.05=7.716375
期望值=7.716375(年)>6(不做手術活6年)

Ans:要

###### tags: `Biomedical Informatics` `110` `2021`

<style>
.navbar-brand::after { content: " × FJUMIIA"; }
</style>