實證醫學Evidence-Based Evidence-Based

Medicine Pharmacy, Medicine Pharmacy, and Nursing and Nursing

20102010

林口長庚紀念醫院 實證醫學中心林口長庚紀念醫院 實證醫學中心風濕過敏免疫科主任風濕過敏免疫科主任 余光輝醫師余光輝醫師

Chief, Center for Evidence-Based MedicineChief, Center for Evidence-Based MedicineChang-Gung Memorial Hospital, TaiwanChang-Gung Memorial Hospital, Taiwan

gout@adm.cgmh.org.twgout@adm.cgmh.org.tw

The Best Evidence Depends The Best Evidence Depends

on the Type of Questionon the Type of Question

What are the phenomena/problems? Observation

What is frequency of the problem? (Frequency) Random (or consecutive) sample

Does this person have the problem? (Diagnosis)

Random (or consecutive) sample with gold standardgold standard Who will get the problem? (Prognosis)

Follow-up of inception cohort How can we alleviate the problem? (Therapy)

Randomized controlled trial (RCT)(RCT)

Current Best EvidenceCurrent Best Evidence

Are the results of the study Validity ( 效度 )

Diagnostic Accuracy Study ( 診斷試驗 )

R: Was the diagnostic test evaluated in an representative spectrum of patients?

mild and severe, early and late, treated and untreated

A: Was the reference standard ascertained regardless of the index test result?

Mbo: Was there an independent, blind comparison with a gold standard of diagnosis?

Was the test validated in a second, independent group of patients?

Importance Diagnostic Test

Accuracy of the test in distinguishing patients with and without the target disorder Sensitivity (Sn) Specificity (Sp) Positive predictive value (PPV) Negative predictive value (NPV) Likelihood ratio (LR)

Diagnostic tests that produce big changes from pre-test to post-test probabilities

Questions to answer in applying a valid diagnostic test to an individual patient

Diagnostic Test “Applicability” Is the diagnostic test available, affordable,

accurate, and precise in our setting? Can we generate a clinically sensible estimate

of our patient’s pre-test probability? From personal experience, prevalence statistics,

practice databases, or primary studies Are the study patients similar to our own Is it unlikely that the disease possibilities have

changed since the evidence was gathered Will the resulting post-test probabilities affect

our management and help our patient? Could it move us across a test-treatment threshold Would our patient be a willing partner in carrying it

out Would the consequences of the test help our patient

reach his or her goals in all this

Critical Appraisal of Diagnostic Accuracy Study

“診斷工具”的評讀

Are the results of the study valid ( 效度如何？ ) Was the diagnostic test evaluated in a representative

spectrum of patients ( 是否經過有代表性的病人群測試過？ ) Was the reference standard ascertained regardless of

the index test result ( 標準診斷工具做確診時不知道指標診斷工具的結果？ )

Was there an independent, blind comparison between the index test and an appropriate gold standard of diagnosis ( 標準診斷與指標診斷工具是在獨立且雙盲的情況下進行比較？ )

What were the results ( 結果是甚麼 ) ？ Are test characteristics presented ( 呈現診斷工具的特性？ )

Can we apply to our patient ( 可以應用到我的病人？ ) Were the methods for performing the test described

in sufficient detail to permit replication?

Was the diagnostic test evaluated in a representative spectrum of patients

是否經過具有代表性的病人群測試過？

最佳狀況為何？ 何處找到相關訊息？指標診斷工具最好經過疾病各層面病人的測試，如不同嚴重程度、不同時期

最好病人亦能隨機選擇或連續性，以減少選擇偏差

“研究方法”應說明如何選入病人，是否隨機選擇，病人的來源，是否具有代表性

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Was the reference standard ascertained regardless of the index test result

標準診斷工具做確診時不知道指標診斷工具的結果？

最佳狀況為何？ 何處找到相關訊息？理想的情況是所有病人都應接受標準診斷工具及指標診斷工具的檢驗

若標準診斷工具是侵入性或昂貴檢查時，則可以選擇指標診斷工具檢查結果陰性者為之或經一段適當時間的追蹤以確定是陰性結果

“研究方法”應說明標準診斷工具是否用於測試所有病人，或是用其他取代的方法，例如指標診斷工具呈陰性患者追蹤一段時間

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Was there an independent, blind comparison between the index test and an appropriate gold standard of diagnosis標準診斷工具與指標診斷工具是在獨立且雙盲的情況下進行比較？

最佳狀況為何？ 何處找到相關訊息？標準診斷工具的選擇是否恰當，有時候單一診斷工具無法做確定診斷，必需結合其他工具以確定疾病的存在

標準診斷工具與指標診斷工具是獨立分開執行且是雙盲的檢查結果的判讀者不應知道另一項檢查的結果為何

“研究方法”應敘述所選用的標準診斷工具為何，必要時應做背景資料的搜尋，看所選擇的標準診斷工具是否恰當

“研究方法”中亦應闡明兩種檢查由誰執行，如何進行，是否獨立雙盲

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Can we apply the results to our patient ？

( 可以應用到我的病人嗎 )

Patients Are your patients similar enough that the

prevalence of the disease in the study population is similar to that in your patients?

Is the severity of the disease in the test population similar to patients you are likely to see?

Benefits Are there risks associated with the tests? Are these outweighed by the danger of an

undiagnosed disease?

評讀有關診斷檢查的研究 (STUDIES

OF DIAGNOSTIC TESTS) 1.   該檢查是否被描述清楚 ? ( 包括被界定異常的點 ) 2.   是否已確立所有病人實際有、無疾病的標準 ( 金字標準 )?

只要能夠把資料填入 2x2 表中，就可以知道該檢查各方面的重要效能。 3.   有、無疾病的受檢病人譜與檢查所要應用的病人之特徵是否配合 ?

敏感度常會受疾病嚴重程度影響，特異度常受研究中無病者特徵的影響。 4.    檢查與疾病狀態的評估是否無偏差 ?

如果是在以之急病狀態下決定檢定結果，可能會出現偏差，反之亦然。 5.     檢查效能是否以敏感度、特異度或概似比來摘要表示 ?

這些是在決定選擇何種檢查方法時所需要的資料。 6.    當檢查值是連續值時，移動分界點對檢查效能有甚麼影響 ?

檢查結果是分界點於哪裡定出正常值與異常值的分界而定。 7.     如果有提及預測值，其臨床上真正的盛行率是否有提出 ?

預測值受盛行率及該檢查的敏感度與特異度有所影響，若有無疾病的受檢者是分開選出，與臨床發生的盛行率無關，則由此計算出來的預測值便無任何臨床意義。

K.H. Yu K.H. Yu TaiwanTaiwan

診斷檢驗Diagnostic testSensitivity. Specificity, PPV, NPV, LR, ROC curve

疾病 疾病 (Disease)

有(Present)

無(Absent)

檢查(Test

)

陽性 (Positive)

真陽性 TPTP a

偽陽性 b FP

陰性 (Negative)

偽陰性 c FN

d 真陰性 TNTN

DiagnosisDiagnostic test (ferritin)

Disease (IDA)

Present Absent

Positive ( 陽性 )

731 TP aa

b FPFP 270

Negative ( 陰性 )

78 FNFN

cd d TN 1500

Sensitivity (Sn) = a/a+c = 731/809 = 90% SnSn SpSpSpecificity (Sp) = d/b+d = 1500/1770 = 85% ( 敏感度會受疾病嚴重程度影響 )

Positive predictive valuePositive predictive value (PPV) (PPV) = = aa//a+ba+b = 731/1001 = 73% = 731/1001 = 73% (=post-test (=post-test probability)probability)

Negative predictive valueNegative predictive value (NPV) (NPV) = = dd//c+dc+d = 1500/1578 = 95%= 1500/1578 = 95%

但 ♠ ♠ 診斷試驗的預測值 (predictive value) 受疾病的盛行率 (prevalence) 影響。Positive predictive value (PPV) = Sen . P / [Sen . P + (1-Sp) . (1-P)] ( 貝氏貝氏定理 )P= 0.5, PPV= 0.8×0.5 / [0.8×0.5+0.2×0.5] = 0.8 = 80.0%P= 0.05, PPV= 0.8×0.05 / [0.8×0.05+0.2×0.05] = 17.4%P= 0.005, PPV= 0.8×0.005 / [0.8×0.005+0.2×0.005] = 0.2%同一診斷工具 , 在不同盛行率情況下 , 其 Predictive value 結果不同。 ~ LRLR 概概似比似比

Specificity 高，但運用在盛行率低的族群時，大部分陽性結果是假陽性。Sensitivity 高，但運用在盛行率高的族群時，大部分陰性結果是假陰性。

PPV PPV aa//a+ba+b

NPV NPV dd//c+dc+d

例： Ovarian cancer CA-125: PPV ~ 2% in screen vs. 97% in pelvic mass cases

LRLR

PrevalencePrevalence ( (different clinical different clinical situationssituations)) affect predictive value

Increasing the Prevalence of Disease Before Testing Increasing the Prevalence of Disease Before Testing : When the prevalenceprevalence of disease in the population tested is relatively high is relatively high － more more than several percent than several percent － the test perform wellthe test perform well.

Diagnostic test

Disease (HIV)(HIV)

Present

Absent

Positive ( 陽性 )

9TP

100 FP

Negative ( 陰性 )

1FN

9890 TN

Community-wide HIV screeningCommunity-wide HIV screening

Test: 90% sensitivity, 99% specific

Population 10,000

Prevalence 0.1% = 10/10,000

PPV = 9/(9+100) = 0.08 = 8%

NPV= 9890/(1+9890) = 0.9999 = 99.99%

例： Ovarian cancer CA-125: PPV ~ 2% in screen vs. 97% in pelvic mass cases

PPVPPV

Note: spectrum of patients, age, gender, risk factors, clinical findings (prevalence)

同一診斷工具 , 在不同盛行率情況下 , 其 Predictive value 結果不同

Increasing the Prevalence of Increasing the Prevalence of Disease Before TestingDisease Before Testing

~~ Prevalence Prevalence affect theaffect the PPVPPV

Test Result

Disease (DM)(DM)

Present Absent

Positive ( 陽性 )

900TP

90FP

Negative ( 陰性 )

100FN

8910TN

Community-wide HIV screeningCommunity-wide HIV screening

Test: 90% sensitivity, 99% specific

Population: 10,000

Prevalence: 0.1% = 10/10,000

PPV = 9/(9+100) = 0.08 = 8%

NPV = 9890/(1+9890) = 0.9999 = 100%

Population = 10,000

Prevalence = 10% = 1000/10,000

Sensitivity = 900/1000 = 90%

Specificity = 8910/9000 = 99%

PPV = 900/990 = 0.91 = 9191%

NPV = 8910/9010 = 0.99 = 99%

Diagnostic test

Disease (HIV)(HIV)

Present Absent

Positive ( 陽性 )

9TP

100 FP

Negative ( 陰性 )

1FN

9890 TN

Diagnosis LRLR: : likelihood ratiolikelihood ratio ((multi-multi-

levels oddslevels odds))

Diagnostic test (ferritin)

Disease (IDA)

Present Absent

Positive ( 陽性 ) 731 TP a

b FPFP(I)

270

Negative ( 陰性 ) 78 FNFN(II) c d TN 1500

Sensitivity = TP/(TP+FN) = TP/(TP+FN) = a/a+c = 731/809 = 90% Sn Sn SpSp SNout, SpPinSNout, SpPin

Specificity = TN/(FP+TN) = TN/(FP+TN) = d/b+d = 1500/1770 = 85%Positive predictive value (PPV)Positive predictive value (PPV) = a/a+b = 731/1001 = = a/a+b = 731/1001 = 73% (=post-test probability)

Negative predictive value (NPV) Negative predictive value (NPV) = TN/(FN+TN) == TN/(FN+TN) = d/c+d = 1500/1578 = 95%d/c+d = 1500/1578 = 95%

LRLR++ for a positive result = Sens/(1- Spec)(1- Spec) = aa/(a+c)/(a+c) / / bb/(b+d)/(b+d) = 90%/15% = 6陽性概似比陽性概似比 LR+ = 敏感度 /(1- 特異度 ) = TP/FP = 有病者與無病健康者 , 檢驗呈陽性的機率勝

算比LRLR-- for a negative result = (1-sens)(1-sens)/spec = cc/(a+c)/(a+c) // dd /(b+d) /(b+d) = 10%/85% = 0.12

Pre-test probability (prevalence)= probability (prevalence)= a+ca+c/a+b+c+d= 31%/a+b+c+d= 31% Pre-test odds = prevalence / (1-prevalence) = 31%/69% = 0.45

** Post-test odds = Post-test odds = Pre-test oddsPre-test odds × × LRLR = 0.45 = 0.45 ×× 6 =2.7 6 =2.7 PPosttest probability = Posttest odds / (odds + 1) = 2.7/(2.7+1) = 73% (= PPV 73%)

+ PV+ PV

- - PVPV

Likelihood Ratio Likelihood Ratio (2, (2, 55, 10, , 10, multi-levels oddsmulti-levels odds))

•LR+LR+ 陽性概似比 (LR of a positive test result, LR+) = 敏感度 / (1- 特異度 )

•LR +LR + = = 有病者與健康者 檢驗呈陽性的機率比 有病者與健康者 檢驗呈陽性的機率比 (( 勝算 勝算 an odd) =an odd) = TP / TP / FPFP

((BenefitBenefit: : multi-levelsmulti-levels, not just , not just binarybinary of yes or no) of yes or no)Diagnostic test (ferritin)

Disease (IDA)

Present

Absent

LR

Very positive < 15 59% 1.1% 52Moderate positive

15-34 22% 4.5% 4.8Neutral 35-64 10% 10% 1Moderate negative

65-94 3.7% 9.5% 0.39Extremely negative

≥ 95 5.9% 75% 0.08

100% 100%

Poor-fair Good Excellent

Positive likelihood ratio

2.1 – 5.0 5.1 – 10.0 > 10

Negative likelihood ratio

0.5 – 0.2 0.19 – 0.1 < 0.1

Diagnosis: Cut-off PointDiagnosis: Cut-off PointTrade-offsTrade-offs between sensitivity and specificitybetween sensitivity and specificity

敏感度與特異度之間的取捨敏感度與特異度之間的取捨

Moving this point changes sensitivity and Moving this point changes sensitivity and specificity of the test. specificity of the test. (trade-offs between Sensitivity and (trade-offs between Sensitivity and Specificity)Specificity)

Receiver operating characteristic (ROCROC) curve ~ selecting a cutoff point for continuous data

IIIIββ

TNFN

TP

Importance of cut-off value on test performance. As the cut-off is moved to left, sensitivity (true positive, TP) increases, but specificity (true negative, TN) decreases. FN = false negative

FP

IIαα

Without Without diseasedisease

With With diseasedisease

Receiver operating characteristic (ROCROC) curve ~ selecting a cutoff point

• 以真陽性 ( 敏感度 ) 為縱軸，假陽性 (1- 特異度 ) 為橫軸，即將有病和沒病健康者試驗結果呈陽性的機率做一比較。

ROC curve: Select a cut-off point for continuous data.

If the area under the ROC curvearea under the ROC curve is 0.50.5 (null hypothesis), the model has no discriminatory power.

• 由檢驗結果，做診斷時 Note: spectrum of patients ( 敏感度會受疾病嚴重程度影響 ), age, gender, risk factors, clinical findings (prevalence)

ROC curveROC curve以真陽性 ( 敏感度 ) 為縱軸，假陽性 (1- 特異度 ) 為橫軸，即將有病和沒病健康者試驗結果呈陽性的機率做一比較。

ROC curve: selecting a cut-off point for continuous data.

(Sensitivity vs. 1-Specificity)

Trade-offs between sensitivity and specificityOptimum cutoff pointcutoff point correlated with the best Youden Youden indexindex = (sensitivity + specificity -1)

TP

FP

UselesUselesss

AUROCAUROC: >0.8: >0.8 0.9 – 1.0 excellent

0.8 – 0.9 good

0.7 -0.8 fair

0.6 – 0.7 poor

< 0.6 useless (null 0.5)

Trade-offs between sensitivity and specificityTrade-offs between sensitivity and specificity

AA

BB

DiagnosisDiagnosis

Use of Sensitivity Test ( 高敏感度檢查的運用 ) Treatable disease Screening: maximize sensitivity while optimizing specificity

未被檢查出來會有嚴重後果者 treatable or transmissible e.g. screen donated blood for HIV, Pap smear, mammograms

Rule out disease ~ (SNout) (SNout) e.g. ANA

Use of Specificity Test ( 高特定度檢查的運用 ) 當假陽性結果會傷害患者身體、情緒、財物時

e.g. cancer chemotherapy Rule in disease ~ (SPin) (SPin) Diagnosis: maximize specificity while optimizing sensitivity. e.g.

anti-ds DNA

ROCROC (receiver operating characteristic) curve 以真陽性 (TP, 敏感度 ) 為縱軸，假陽性 (FP, 1- 特異度 ) 為橫軸，即將有病及無病

者檢驗結果呈陽性的機率做一比較。 ROC curve 下方的面積越大，診斷工具的準確準確度度越好

Optimum cutoff pointcutoff point correlated with the best Youden index = (sensitivity + specificity - 1) c.f. LR+c.f. LR+ for a positive result = sensitivity / (1- specificity): The probability of that test result in people

with the disease divided by the probability of the result in people without disease Accuracy: proportion of correct results = (a+d) / (a+b+c+d) = (prevalence x sensitivity) + (1-prevalence) x

(specificity) = TP+TN / (TP+TN+FP+FN)

Diagnosis Strategies Diagnosis Strategies Serial vs. parallel testSerial vs. parallel test

Serial Test ( 高特定度 ) The result of test 1 are considered before test 2, and so on In order to be considered positive, all test in the series must be

positive Highly specific but insensitive

Useful when false positive are undesirable Such as treatment is highly invasive or toxic e.g. cancer

chemotherapy

Parallel Test ( 高敏感度 ) Any positive is considered a positive

Sensitive but not specific Useful when rapid diagnosis is necessary and a missed

diagnosis in undesirable

Take Home MessageTake Home Message**

實證醫學五步驟 EBM 步驟一：形成一個可以回答的臨床問題

((Ask)Ask)

試著將您的問題分成下列四個部分（ PICO ）：

□ Patient or Problem 病人或問題□ Intervention or Indicator 介入或指標

某種治療、檢查、危險因子等□ Comparator 比較－ 該治療和甚麼相比？□ Outcome 結果－ 您想要達成或避免甚麼？

* Question Log – A tool for just in time learning. Carl Heneghan, Paul Glasziou 以下“ Take home message” 資料主要係參考由台北醫學大學．市立萬芳醫院實證醫學中心經授權作中文編譯，英國 The Centre for Evidence-Based Medicine 及 The National Library for Health, NHS 所出版由Carl Heneghan 及 Paul Glasziou 撰寫之 Questions Log: A tool for ‘just in time’ learning 手冊。任何形式之引用請註明文獻出處。

EBM 步驟二：搜尋最佳證據 (Acquire)

利用問題的 PICO 結構（如上述 Ask ）設定搜尋策略。

回想問題中各 PICO 部份的每一個辭彙及同義詞。一次就單一 PICO 元素進行搜尋。如，從介入（ Intervention ）開始，但必須確定你已聯集（ OR ）所有的同義詞。 可以使用截斷字 (truncation) ，並加上” *” ，如以

child* 取代 children 搜尋文件：請試著從 Cochrane 開始；其他問題型態則建議試試 PubMed: Clinical Queries 或 National Library for Health (NLH) 。

Dr. Yu KH ~ PubMed 快速搜尋技巧： (P) and (I) and (Cochrane or meta-analysis or systematic review)

步驟三 Appraisal ：嚴格評讀證據之(a) 效度 與 (b) 重要性 ( 效益大小 )

(a) 效度 (Validity) 各種形式的問題都含以下三個共同項目 (RAM-bo) 研究族群是否具有代表性 (Representative)?

隨機選擇（ random selection ）／連貫性／起始點病人群，或者如果是比較性的研究，組別間是否可以比較？隨機分派（ random allocation ）／調整

是否有足夠的確認和追蹤 (Ascertainment/follow-up)? 反應率／追蹤／確認＞ 80%

結果的估計值測量 (Measurement) 是否公正無偏？恰當？ 結果以盲法（ blinded ）或客觀的（ objective ）估計 以上這些答案，通常可以在文章中的方法學（ Method ）部分和結果（ Result ）的第一、二段中找到。這樣的評讀，一開始可能會令您覺得困難重重（就像騎腳踏車一樣），但是，累積了一些經驗之後，您只要幾分鐘就能完成

(b) 重要性－效益大小 看結果段 (Results section) 中描述的主要結果。效果有多大？多重要？統計意

義要看信賴區間及 P值；相對危險（ relative risk ）、相對危險性降低度（ relative risk reduction ）、勝算比（ odds ratio ）代表生物學上的影響。效果的絕對估計值：絕對危險性降低（ absolute risk reduction ）、益一需治數（ NNT, number needed to treat ）則代表在臨床上對病人的影響。

EBM 步驟四：將證據與臨床專業經驗及病人期望結合 (Apply)

□ 您的病人是否與研究中的病人差別很大，以至 於無法適用該研究結果？

□ 您期望您的病人從研究結果中獲得多大的好處？

□ 還有哪些替代方案？□ 研究結果適用於您的病人嗎？□ 病人的想法為何？

研究效果需要因應個別病人做調整，如治療 Patient NNT = 1/ (RRR × PEER)

EBM 步驟五：評估執行效果及效用－勤做紀錄，改善過程

最後一個步驟，看在執行過程中，您的表現如何？您可能要問自己下列幾個問題：

□ 您正在紀錄您的問題嗎？□ 您是否正在廣大的資源中尋找有用的外部證據？□ 您有能力將這些證據應用在適當的病人身上嗎？□ 您是否依循這些新證據來改變您的診療習慣？

第十屆醫療品質獎

參 考 附 件 參 考 附 件 AppendixAppendix

For further reading A: critical appraisal checklist B: statistical concepts

評讀評讀 附件附件 : : Critical Critical AppraisalAppraisal

OxfordOxford - Critical Appraisal Sheets 1. RCT (therapy study) 2. Systematic review 3. Diagnostic test

CASPCASP Critical Appraisal Skills Programme (CASP)

appraisal tool http://www.phru.nhs.uk/pages/PHD/resources.htm

推薦中文版 Clinical Epidemiology: The Essentials. Third edition. Robert H. Fletcher, et al. Wagner. Williams & Wilkins. 1996

決定臨床研究效度的基本指引所有研究 (ALL STUDIES)

1.  本研究想要回答何種臨床問題 ? 研究設計應配合臨床問題

2.   研究的病人、變項及結果為何 ? 這幾點決定研究結果的可類推性

3.   研究結果是由偏差造成的可能性有多大 ? 組間若有系統差異 ( 譬如病人特徵、介入、危險因子、結果、測量方

法等 ) ，將會降低內在效度。 4.  效果有多大 ?

臨床決定不只考慮有無效果，還要考慮效果的大小。 5.   研究結果是由機會造成的可能性有多大 ?

須知道真實效果可能出現的範圍 (信賴區間 ) ，及 ( 用處較小 )觀察結果是單由機會造成的可能 (「陽性」結果用 p值，「 陰性」 結果用檢力 ) 。

Critical Appraisal of Therapy Study“治療研究”的評讀

Are the results of the trial valid ( 效度如何 ) ？ Was the assignment of patients to treatment

randomized ( 是隨機分配嗎？ ) Were the groups similar at the start of the trial ( 試

驗開始時兩組條件是否相似？–表一 ) Aside from the allocated treatment, were groups

treated equally ( 兩組的其他治療一樣嗎？ ) Were all patients who entered the trial accounted

for? - and were they analyzed in the groups to which they were randomized ( 所有進入試驗者皆列入統計，並依所分配的組別計算？ ITT analysis)

Were measures objective or were the patients and clinicians keep blinded to the treatment ( 結果的測量客觀？受試者及醫師都不知道所接受的治療為何？盲性評估 )

What were the results ( 結果為何 ) ？ (important?)

How large was the treatment effect (治療的效果有多大？ )

RR (relative risk) ARR (absolute risk reduction) RRR (relative risk reduction) NNT (number needed to treat)

How precise was the estimate of the treatment effect (治療效果的預測有多準確？ )

Point estimate 點估計 CI (confidence interval) 95%信賴區間 (cover null hypothesis?)

Will the results help me in caring for my patient ( 結果適用於我的病人嗎 ) ？ (applicability?)

Critical Appraisal of Therapy Study“治療研究”的評讀

Was the assignment of patients to treatment randomized

是隨機分配嗎？

最佳狀況為何？ 何處找到相關訊息？理想狀況是以中央電腦隨機分配，尤其是跨中心的研究

較小型臨床試驗可以獨立人士如醫院藥局“執行”隨機分配　　　　

“研究方法”應敘述受試者如何做分配，以及是否隱匿式隨機分配 (randomization, and allocation concealment)

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were the groups similar at the start of the trial

試驗開始時兩組條件是否相似？

最佳況為何？ 何處找到相關訊息？如果隨機分配的方式及過程正確，實驗組與對照組應該相似

應有統計比較兩組是否有差異　　　

“研究結果”的表格比較兩組之間可能影響研究結果的基本條件及風險因子的差異，如年齡、性別、危險因子等(通常在 Result Table 1 p 值 )

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Aside from the allocated treatment, were groups treated

equally兩組其他的治療條件一樣？

最佳狀況為何？ 何處找到相關訊息？除了要研究的治療方式或檢驗外，兩組病人所接受的其它治療都要一樣

“研究方法”中所敘述的追蹤計劃，及允許的其它治療

“研究結果”中兩組實際接受各項額外治療的種類與比例

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were all patients who entered the trial accounted for and were they analyzed in the groups to which they were randomized

所有進入試驗者皆列入統計，並依所分配的組別計算？

最佳狀況為何？ 何處找到相關訊息？追蹤過程中，病人流失越少越好，流失率應低於 20%達到研究結果的病人數太少時，即使流失率低，研究結論仍可能有偏差統計分析時應按照病人隨機分配的組別計算 (Intention-to-treat analysis)

“研究結果”中應說明多少病人接受隨機分配 (Table 1)，多少病人進入統計分析。病人流失的數目及原因。

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were measures objective or were the patients and clinicians were blinded結果的測量客觀，受試者及醫師都不知道所接受的治療為何？

最佳狀況為何？ 何處找到相關訊息？雙盲試驗：受試者與研究者皆不知道受試者所接受的治療為何

結果評估的客觀性：結果的評估若是客觀的（如死亡），則較不影響，但若結果的判斷是主觀的（如症狀、功能、能力）則盲性評估就很重要

“研究方法”應說明對照組如何做偽裝治療，如給予外觀一樣的安慰劑

“研究方法”應說明結果如何做評估，評估者是否知道病人接受的治療為何

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

How large was the treatment effect

治療效果有多大？

呈現的方式 代表的意義Relative Risk ( 相對風險 ) (RR)

RR = 0.10 / 0.15 = 0.67

治療組發生風險相對於對照組的倍數。 RR=1 兩組無差別，RR<1 治療可降低風險， RR>1 治療會增加風險RR<1 表示治療可降低死亡的風險

Absolute Risk Reduction (ARR)

(絕對危險性降低 )

ARR = 0.15 – 0.10 = 0.05 or 5%

治療組與對照組發生風險的絕對差異

治療的益處是降低 5% 的死亡率

Relative Risk Reduction (RRR)

( 相對風險性降低 )

RRR = 0.05 / 0.15 = 0.33 or 33%

or RRR = 1 – 0.67 = 0.33 or 33%

相對於對照組，治療組降低風險的比率 ( 最常見的呈現方式 )

相對於對照組，治療可以降低死亡的的機率 33%

Number Needed to Treat (NNT)

(益一需治數 )

NNT = 1 / ARR = 1 / 0.05 = 20

要預防一位病人不良結果發生，所必需治療的病人數

治療 20位病人 2 年，才能預防 1 人死亡

某研究追蹤二年，對照組死亡率 15%，治療組死亡率 10%，結果呈現的方式有下列幾種：

How precise was the estimate of the treatment effect

治療效果的預測多準確？

The true risk of the outcome in the population is not known The best we can do is to estimate the true risk based on the

sample of patients in the trial We can gauge how close this estimate is to the true value by

looking at the confidence intervals (CI) Narrow CI represents a precise reflection of the population

value The CI also provides us with information about the statistical

significance of the result If the value corresponding to no effect falls outside the 95% CI then the

result is statistically significant at the 0.05 level If the CI includes the value corresponding to no effect then the results

are not statistically significant (null hypothesis) ITT vs. per protocol analysis: 如果不是每位病人都實際接受被分派的治療，這時必須依不同目的與不同科學強度進行不同的分析。『意圖治療』分析是針對管理決策，所以是以分派的治療來分析；『 per protocol』分析是要解釋介入本身的效果，所以要以實際接受的治療分析。

Will the Results Help Me in Caring for My Patients ?

Is our patient so different from those in the study that its results cannot apply ? (Are the people in the study like my patient? – note inclusion criteria)

Data from Taiwan, China, or Asia ( 種族差異 )? Cost-effectiveness analysis ( 效益分析 ) Do I miss any data? 同時手動搜尋所選文章之參考文獻及專家回顧文獻 與詢問專家 Age, general state of health, type and severity of disease process,

time in the course of the disease (i.e. applicability) Is the treatment feasible in my setting? Will the potential benefits of treatment outweigh

the potential harms of treatment for my patients? Did the study cover all aspects of problem? Does it suggest a clear and useful plan of action?

help to clarify a patient’s prognosis suggest a useful plan to improve patient’s state of health

統合分析 (META-ANALYSES) 1.  是否已找出所有相關的研究

包括已出版及未出版的研究 ? 目的是要綜合全體研究的結果，而非有偏差的樣本研究。

2.  是否只包含合乎科學原則的研究 極少偏差的研究 ?

研究必須依據最可靠的證據。 3.   估計效果時 :

a.  研究是否具同性質 ( 病人、介入及研究結果均類似 )? 從完全不相似的研究中找出整體效果的測量值是不合適的。

b.   是否以各研究的樣本大小加權計算 (weighting)? 加權計算時，樣本數大 (精密度高 ) 者的權值是否較樣本數小 (精密度低 )

者為大 ? 4.   研究品質與結果有無相關 ?

品質佳的研究較可信。

Critical Appraisal of Systematic Review

“系統性回顧”的評讀 Are the results of the review valid ( 效度如何 ?)

What question (PICO) did the systematic review address (回答什麼問題 ?)

Is it unlikely that important, relevant studies were missed ( 有沒有遺漏重要的文獻?)

Were the criteria used to select articles for inclusion appropriate (選擇文獻的條件準則適當 ?)

Were the included studies sufficiently valid for the type of question asked (選擇的文獻有效回答所問的問題 ?)

Were the results similar from study to study ( 各研究間結果相似 ? 無異

質性 ) What were the results ( 結果為何 ?)

How are the results presented ( 結果如何呈現 ? Meta-analysis Forest plot, heterogeneity chi-square Cochran Q)

What question (PICO) did the systematic review addressed

系統性回顧想要回答什麼問題？

最佳狀況為何？ 何處找到相關訊息？

應清楚闡明文章想要回答的問題，暴露因子 (包括治療、檢驗等 )與結果的因果關係簡單明瞭

題目、摘要或前言的末段應清楚描述所關心的問題

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Is it unlikely that important, relevant studies were missed

沒有遺漏重要的文獻？

最佳狀況為何？ 何處找到相關訊息？資料搜尋是否完整，包含 -- 重要資料庫如Medline,Cochrane,EMBASE -- 相關研究的參考文獻 -- 向專家請教，特別是尚未刊載的研究 -- 不只限於英文資料 -- 搜尋策略包括 MESH term 及 text words

“研究方法”詳細描述搜尋字彙與策略

“研究結果”詳列回顧的題　目、摘要、全文數目，排　除的文章及排除理由，並　以圖表或流程圖呈現

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were the criteria used to select articles for inclusion appropriate

選擇文獻的準則適當 ?

最佳狀況為何？ 何處找到相關訊息？事先清楚界定“選入”及“排除”文章的準則

準則的描述應包括病人的特性、介入治療的方法或暴露因子、有興趣的研究結果、及研究設計的類型

“研究方法”詳細描述文章“選入”及“排除” 的準則， 通常也包含研究的類型

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were the included studies sufficiently valid for the type of question asked

選擇的文獻有效回答所問的問題 ?

最佳狀況為何？ 何處找到相關訊息？應描述所回顧的每篇研究文章的品質，研究品質的判定準則依不同臨床問題而事先擬定，如隨機分配、雙盲、追蹤的完整等　　　　

“研究方法”應描述品質的評估及所使用的準則

“研究結果”應說明各研究的品質

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Were the results similar from study to study

各研究的結果相似 ?

最佳狀況為何？ 何處找到相關訊息？最理想的狀況是各研究的結論一致或同質性。如果各研究的結果有異質性，作者可估計差異是否顯著﹝卡方檢定﹞，並且探討各研究間異質性的原因　　

“研究結果”應說明各研究的結果是否有異質性，並討論可能的原因Forest plot應顯示異質性的卡方檢定結果，若有異質性，應討論引起異質性的原因

□ 是　　　　　　　□ 否　　　　　　　□ 不清楚

評論：＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿ ＿＿＿

Meta-analysis ( 統合分析 ) If the results of the individual studies are similar Gives weighted values to each of the individual studies according to their

size The individual results of the studies expressed in a standard way, e.g.

relative risk, odds ratio or mean difference between the groups Results are displayed in a figure called “forest plot”

Exploring heterogeneity ( 檢驗異質性 ) Investigated by “visual examination” Cochran chi-square ( Cochran Q )

Definite heterogeneity ( 確定有差異 ) Cochran Q ( p< 0.1 )

Possible heterogeneity ( 可能有差異 ) Cochran Q is not statistically significant Cochran Q / degrees of freedom (Q/df) > 1

Heterogeneity unlikely ( 有差異機會不大 ) Cochran Q is not statistically significant Q/df < 1

How are the results presented 結果如何呈現 ?

盛行率研究 (PREVALENCE

STUDIES)

1.   定義病例的準則為何 ? 要能明確定義何者為病例才能計算盛行率。

2.   病例在甚麼族群中被發現 ? 盛行率受到所描述的族群之特徵影響。

3.   盛行率的描述是否來自族群中的無偏差樣本 ? 以樣本盛行來估計族群盛行率時，樣本必須是無偏差的。

世代研究 (COHORT STUDIES)

1. 世代的所有成員 a.  是否於追蹤開始時加入 (起始世代 )？

若不是這樣，世代成員極好或極壞的表現便不能納入研究結果內。 b.  是否都有可能發生該疾病 ?

追蹤已罹患此病及不可能發生的病人之結果均無意義。 c.  是否都是在疾病進程的相似點開始觀察 (零時 )?

若在疾病進程的不同點開始觀察並計算結果事件，所獲得的預後會不同。 2.  是否對所有成員進行完整的追蹤 ?

中途脫離世代者如果平均而言比留在世代者有較佳或較差病情時，研究結果可能會因此產生偏差。

3. 世代的所有成員是否都採用同一強度來評估病情 ? 如果不是的話，研究結果的差異可能來自測量偏差而非真正的差異。

4.   是否屬無偏差比較 ? (除了研究變項外，世代成員在其他可影響結果的特徵上是否相同 ?)

為了將結果歸因為研究因子的作用，比較組間的其他影響因子必須平均分布。

病例對照研究(CASE-CONTROL STUDIES)

1.   病例是否在疾病發作時即納入研究 ? 現存病例的危險因子可能會受發病期長短所影響。

2.  除了暴露因子外，病例組與對照組是否十分相似 ? 在無偏差的比較過程中，才能有效估計其相對危險。

3.   是否對病例組與對照組採用類似且無偏差的測量 ? 暴露的測量若有偏差，會錯誤估計相對危險的真實值。

Is the evidence about prognosis valid ?

預後研究效度如何 Was a defined, representative sample of patients

assembled at a common (usually early) point in the course of their disease? 病人代表性是否足夠

Was patient follow-up sufficiently long and complete? 追蹤是否完整，時間夠長 (> 80% follow up rate)

“5” and “20” rule Were objective outcome criteria applied in a “blind”

fashion? 結果評估是否以盲法進行 If subgroups with different prognosis are identified?

Was there adjustment for important prognostic factors? Was there validation in an independent group of “test-set”

patients? 分組結果不同時是否考慮到一些重要預後因子的校正？是否經另一組病人的驗證？