8.3 🩺 內科專科考前版
對象:R2-R3 / Fellow,準備台灣內科專科考。本章 specialist 重點:EBM PE deep dive、POCUS、HCMS、AI 整合。
8.3.0.1 📌 一頁重點整理
- 22E 強調 EBM PE 仍是 internist 核心技能:McGee + Simel JAMA Rational Clinical Exam series 是 reference
- POCUS(Point-of-Care Ultrasound) 是 PE 的延伸 tool
- AI alpha-validation for PE 仍 emerging(imaging 領先 PE)
- 📍 台灣:JCT (台灣專科醫師考) PE OSCE 站每年改進
- Verghese 派強調 ritual + humanism;對 EHR 過度依賴持續 critique
8.3.0.2 📜 22E vs 21E 主要差異
| 項目 | 21E | 22E |
|---|---|---|
| POCUS integration | 略 | 強化(Verghese 持續推) |
| AI in PE | 無 | 提及(auscultation AI、retinal scan) |
| EHR copy-paste 警告 | 略 | 強化 |
| HINTS battery | 提 | 詳細 |
| EBM PE evidence | McGee 5e | 22E 引述 5e 與後續 update |
8.3.0.3 🧠 深度概念
8.3.0.3.1 Evidence-Based Physical Examination(EBPE)
核心 principles: 1. PE finding 都該有 sensitivity / specificity / LR 數據 2. 多數傳統 finding LR 介於 0.5-2(useless),少數 highly informative 3. McGee 的 Evidence-Based Physical Diagnosis 是 gold-standard reference 4. JAMA Rational Clinical Exam series(1998-)是 evidence base
8.3.0.3.2 Reliability vs Accuracy
Reliability (κ):兩個 observer 是否一致 - 不同 finding 差異大(abdominojugular 0.92 vs neck vein 0.08-0.71)
Accuracy (LR):finding 是否真反映疾病 - 反映 disease prevalence + signal strength
→ 高 reliability + 高 accuracy = 「useful in clinic」
8.3.0.3.3 POCUS(Point-of-Care Ultrasound)
22E 提到 POCUS 整合: - 已是 ICU、ED、internal medicine 常規 - 例:FAST exam、cardiac PoCUS、IVC、lung ultrasound、DVT bedside US - 限制:operator-dependent、limited training in many programs - 內科 fellow 越來越要求 POCUS competency
8.3.0.3.4 AI in PE Augmentation
22E 略提: - Auscultation AI(電子聽診器):增強 murmur、wheeze 識別 - Retinal AI(fundus photo):DR、glaucoma、CV risk - Voice analysis:detect Parkinson、dementia - Wearable:HR variability、AF - 限制:bias、generalizability
8.3.0.3.5 Diagnostic Imperatives 對應 PE
某些「不能漏」的 dx 對應 PE finding: - Aortic dissection:BP differential 雙手、pulse deficit - SAH:meningismus、neuro deficit - Meningitis:Kernig、Brudzinski(雖 sensitivity 低) - Stroke:HINTS for posterior、NIHSS for cortical - Cardiac tamponade:pulsus paradoxus、Beck’s triad - PE:Wells score with PE-specific findings
8.3.0.4 🩺 高 yield PE Manoeuvres(22E 列出 + LR)
8.3.0.4.1 Heart Failure (Elevated Left Heart Filling Pressure)
- Positive abdominojugular reflux: LR 8.0(最強)
- Displaced apical impulse: LR 5.8
- HR > 100: LR 5.5
- S3 gallop: LR 3.9
- → 多 finding 組合提升 yield
8.3.0.4.2 Pneumonia
- Egophony: LR 4.1
- Percussion dullness: LR 3.6
- Bronchial breath sounds: LR 3.3
- Crackles: LR 2.8
8.3.0.4.3 Ascites(abdominal distension)
- Fluid wave: LR 5.0
- Edema present: LR 3.8
- Shifting dullness: LR 2.3
- Edema absent: LR 0.2(rule out)
8.3.0.4.4 Hepatopulmonary Syndrome
- Clubbing: LR 4.3
- Cyanosis: LR 4.4
- → Combined → high probability in cirrhotic
8.3.0.4.5 Bedside Stop Rules
| Rule | 用途 | Negative LR |
|---|---|---|
| Ottawa Ankle | 排 ankle/foot fracture | 0.1 |
| Wells DVT + neg D-dimer | 排 DVT | 0.2 |
| HINTS peripheral | 排 posterior stroke | 0.02 |
| Alvarado ≤ 4 | 排 appendicitis | 0.1 |
| Heckerling 0-1 | 排 pneumonia | 0.3 |
| Probe-to-bone neg | 排 osteomyelitis (DM foot) | 0.2 |
8.3.0.4.6 HINTS Battery 詳解
Step 1 - Head Impulse Test: - Examiner rotates head while patient fixates - Peripheral:corrective saccade(眼跟不上 → 瞄向回來) - Central:no saccade - → Peripheral pattern in head impulse 是 reassuring(與其他 vertigo 直覺反過來!)
Step 2 - Nystagmus pattern: - Peripheral:unidirectional - Central:direction-changing in different gaze positions
Step 3 - Test of Skew (Cover-uncover): - Negative:no vertical realignment - Positive:vertical correction → central
→ 三項全「peripheral」(head impulse + unidirectional nystagmus + no skew)→ posterior stroke LR 0.02
→ 三項任一「central」→ MRI 必做
8.3.0.5 🌟 Clinical Pearls (8 條)
- Combine multiple PE findings:single LR 有限;combined LR can rule in/out 強烈
- 「Negative finding」也是訊息:edema 不存在 LR 0.2 for ascites
- HINTS 比 NIHSS 對 acute vertigo 更敏感於 stroke:但需 ENT-trained 操作
- Abdominojugular reflux 是 underused gem:bedside 1 分鐘 LR 8.0 for HF
- 「Auscultatory percussion」應淘汰:低 reliability
- POCUS 是 PE 的 21 世紀延伸:不是替代,是 augmentation
- EHR copy-paste 是 PE 偷懶共犯:cut-and-paste 假成 thorough exam
- Verghese 學派 + Feinstein 學派 = EBPE 雙基:humanism + evidence
8.3.0.6 🔍 特殊情境
8.3.0.6.1 1. Verghese vs Technologist 二派
- Verghese: ritual、humanism、bedside teaching
- Tech-heavy: imaging-first、AI-augmented
- 22E 立場:兩者結合(不是對立)
8.3.0.6.2 2. POCUS 在 internal medicine
- Cardiac POCUS:LV function、effusion、IVC
- Pulmonary POCUS:B-lines、pneumothorax、pleural effusion
- DVT bedside US:2-point compression
- AAA screening
- 限制:training、credentialing、liability
8.3.0.6.3 3. 教育議題:bedside teaching erosion
- Trainees 多在 conference room round
- 病人 bedside 互動減少
- 22E 強調必須 reverse this trend
- 解法:dedicated bedside teaching、structured PE curricula
8.3.0.6.4 4. PE 在 telehealth 時代
- Limitations:not hands-on、limited inspection
- Adaptations:patient self-PE guidance、family member assist、video assessment
- 仍需 in-person follow-up for 重要 PE
8.3.0.7 🔬 22E 提到的新進展
8.3.0.8 📍 台灣 Context 專區
8.3.0.8.1 台灣 PE 訓練
- 醫學系課程:preclinical 含 physical diagnosis
- PGY OSCE 站:基本 PE
- 內科專科考:含 PE site
- 但 bedside teaching 持續弱化(與美國類似)
8.3.0.9 ⚠️ 老闆地雷區
- EHR copy-paste 假裝 complete PE
- 不做 bedside US 即使 indicated
- HINTS 不會用 → over-MRI
- OG vs PE conflict 時:不檢視自己 PE 而盲信 imaging
- Bedside teaching 跳過:trainee skill 退化
- Bedside US 沒 credentialing 自做
- 病人 disrobed 不充分:missed dermatologic finding
- Telehealth 過度信任:missed hands-on signs
8.3.0.10 🎓 內科專科考重點預測
8.3.0.10.1 高機率題型
- HINTS battery 內涵與 LR
- HF / pneumonia / ascites LR table 解讀
- Bedside rules(Ottawa、Wells、Alvarado)
- κ scale 解讀
- Pretest probability 應用
8.3.0.11 📖 延伸閱讀
- McGee S. Evidence-Based Physical Diagnosis, 5th ed. Elsevier, 2022.
- Simel D, Rennie D, eds. The Rational Clinical Examination: Evidence-Based Clinical Diagnosis. JAMA Press.
- Kattah JC et al. HINTS to diagnose stroke. Stroke 40:3504, 2009.
- Verghese A, Charlton B et al. Bedside in the modern era. Lancet 391:e23, 2018.
- 中文:McGee 中文版
8.3.0.12 📚 三階段教材索引
- 醫學生概念 → Ch 8 medstudent.md
- 國考衝刺 → Ch 8 board-prep.md
- Harrison 22E 原文 → Ch 8
⚠️ 本 md 為 claude-opus-4-7 撰寫(2026-05-07),未經盧醫師驗證。