循證臨床實(shí)踐00課件

1、循證臨床實(shí)踐Evidence based Practice福建醫(yī)科大學(xué)附屬第一醫(yī)院 陳君敏(drjunminchen)1課程介紹 循證臨床實(shí)踐重點(diǎn)介紹臨床循證實(shí)踐的技能和方法如何針對(duì)臨床遇到的疑難問(wèn)題查找回答問(wèn)題的臨床研究文獻(xiàn)如何分析解讀臨床研究文獻(xiàn)報(bào)告的結(jié)果如何將這些研究結(jié)果應(yīng)用于臨床實(shí)踐如何開(kāi)展證據(jù)研究課程旨在幫助臨床醫(yī)生應(yīng)用當(dāng)前最佳的研究證據(jù)提高診療水平。2概 論循證醫(yī)學(xué)的概念循證醫(yī)學(xué)的起源循證醫(yī)學(xué)的概念循證臨床實(shí)踐3隨著醫(yī)學(xué)科學(xué)的迅速發(fā)展，臨床醫(yī)學(xué)的研究也非?；钴S，每天都有許多醫(yī)學(xué)論文發(fā)表，有新的科學(xué)證據(jù)產(chǎn)生。這些新的成果只有及時(shí)被臨床醫(yī)生了解和應(yīng)用，病人才能得到最好的治療。醫(yī)學(xué)研究臨

2、床應(yīng)用4臨床應(yīng)用嚴(yán)重滯后于醫(yī)學(xué)研究有充分證據(jù)證明有效的治療方法長(zhǎng)期未被使用如心肌梗死的溶栓療法在20世紀(jì)70年代已有多篇文獻(xiàn)證實(shí)，在80年代末才得以廣泛應(yīng)用研究證明無(wú)效，甚至有害的治療方法卻長(zhǎng)期應(yīng)用如利多卡因預(yù)防心肌梗死后的心律失常57循證醫(yī)學(xué)的提出我們生活在知識(shí)爆炸的時(shí)代，醫(yī)學(xué)研究的信息呈指數(shù)增長(zhǎng)。醫(yī)生要保持自己的知識(shí)更新非常困難，而且越來(lái)越困難。 醫(yī)療資源有限，合理使用有限的醫(yī)療資源是各國(guó)的普遍要求 8什么是循證醫(yī)學(xué)？Evidence based practice is the conscientious, explicit and judicious use of current bes

3、t evidence in making decisions about the care of individual patients.慎重、準(zhǔn)確和明智地應(yīng)用當(dāng)前所能獲得的最好研究依據(jù)制定出病人的治療措施 David Sackett 1011傳統(tǒng)醫(yī)學(xué)：以經(jīng)驗(yàn)醫(yī)學(xué)為主，即根據(jù)醫(yī)師的經(jīng)驗(yàn)直覺(jué)來(lái)處理病人，這些經(jīng)驗(yàn)可能來(lái)源于：個(gè)人的臨床實(shí)踐積累專業(yè)培訓(xùn)老師傳授大多數(shù)同行的診療方法 教科書(shū)專家意見(jiàn)遵循證據(jù)是EBM的核心思想什么是證據(jù)？12Patient preferenceResearch evidence14循證醫(yī)學(xué)概論概念（小結(jié)）EBM起源概念（慎重、準(zhǔn)確和明智地應(yīng)用當(dāng)前所能獲得的最好研究依據(jù)制定

4、出病人的治療措施 ）核心思想（遵循證據(jù)）循證臨床實(shí)踐（EBCP）和廣義的EBMEBCP三要素 （臨床經(jīng)驗(yàn)、臨床證據(jù)、患者意愿）15醫(yī)學(xué)證據(jù)醫(yī)療決策的依據(jù)依據(jù)的優(yōu)劣決定決策的成敗17醫(yī)療決策 懷疑病人患某種疾病，醫(yī)生要決定讓病人做某項(xiàng)檢查以確診或排除這種疾病 采用這種檢查而不用另一種檢查 ？病人出現(xiàn)某種臨床問(wèn)題，醫(yī)生要為病人選擇某種治療措施 采用這種治療而不用另一種治療？18醫(yī)學(xué)證據(jù)這種檢查比另一種檢查更準(zhǔn)確、對(duì)病人傷害小、價(jià)格便宜，等等 何以證明？這種治療措施比另外一種效果更好、副作用較小、以及花費(fèi)少 何以證明？19能證明嗎？醫(yī)生的經(jīng)驗(yàn)直覺(jué) 醫(yī)學(xué)理論（比如中醫(yī)的某種學(xué)說(shuō)）推理 科學(xué)實(shí)驗(yàn)（比如動(dòng)

5、物實(shí)驗(yàn)和體外實(shí)驗(yàn)）？教科書(shū)？專家意見(jiàn) 20臨床研究作為醫(yī)學(xué)證據(jù)（按研究?jī)?nèi)容分類）關(guān)于病因的臨床研究關(guān)于診斷或篩查的臨床研究關(guān)于治療或干預(yù)的臨床研究關(guān)于預(yù)后的臨床研究21證據(jù)的分類原始研究證據(jù)觀察性研究：隊(duì)列研究，病例-對(duì)照研究，橫斷面調(diào)查，描述性研究，病例系列，個(gè)案報(bào)告試驗(yàn)性研究：隨機(jī)對(duì)照試驗(yàn)（RCT），非隨機(jī)同期對(duì)照試驗(yàn)，交叉試驗(yàn)，前后對(duì)照試驗(yàn)二次研究證據(jù)：系統(tǒng)評(píng)價(jià)（SR），臨床實(shí)踐指南，臨床決策分析，衛(wèi)生技術(shù)評(píng)估，衛(wèi)生經(jīng)濟(jì)學(xué)研究2224251.Clinical study of recombinant adenovirus-p53 combined with fractionated st

6、ereotactic radiotherapy for hepatocellular carcinoma.Yang ZX, Wang D, Wang G, Zhang QH, Liu JM, Peng P, Liu XH.J Cancer Res Clin Oncol. 2010 Apr;136(4):625-30. Epub 2009 Oct 31.PMID: 19882171 PubMed - indexed for MEDLINERelated citations2.An update on gene therapy in China.Shi J, Zheng D.Curr Opin M

7、ol Ther. 2009 Oct;11(5):547-53. Review.PMID: 19806503 PubMed - indexed for MEDLINERelated citations3.Phase I study of repeated intraepithelial delivery of adenoviral p53 in patients with dysplastic oral leukoplakia.Zhang S, Li Y, Li L, Zhang Y, Gao N, Zhang Z, Zhao H.J Oral Maxillofac Surg. 2009 May

8、;67(5):1074-82.PMID: 19375021 PubMed - indexed for MEDLINERelated citations4.A patient with huge hepatocellular carcinoma who had a complete clinical response to p53 gene combined with chemotherapy and transcatheter arterial chemoembolization.Tian G, Liu J, Sui J.Anticancer Drugs. 2009 Jun;20(5):403

9、-7.PMID: 19318914 PubMed - indexed for MEDLINERelated citations5.The application of gene therapy in China.Peng Z, Yu Q, Bao L.IDrugs. 2008 May;11(5):346-50. Review.PMID: 18465676 PubMed - indexed for MEDLINERelated citations6.Recombinant adenovirus-p53 gene therapy combined with radiotherapy for hea

10、d and neck squamous-cell carcinomaZhang SW, Xiao SW, Liu CQ, Sun Y, Su X, Li DM, Xu G, Zhu GY, Xu B.Zhonghua Zhong Liu Za Zhi. 2005 Jul;27(7):426-8. Chinese. PMID: 16188130 PubMed - indexed for MEDLINERelated citations7.Current status of gendicine in China: recombinant human Ad-p53 agent for treatme

11、nt of cancers.Peng Z.Hum Gene Ther. 2005 Sep;16(9):1016-27. Review. No abstract available. PMID: 16149900 PubMed - indexed for MEDLINERelated citations8.Gendicine: the first commercial gene therapy product.Wilson JM.Hum Gene Ther. 2005 Sep;16(9):1014-5. Chinese, English. No abstract available. PMID:

12、 16149899 PubMed - indexed for MEDLINERelated citations9.DNA-based therapeutics and DNA delivery systems: a comprehensive review.Patil SD, Rhodes DG, Burgess DJ.AAPS J. 2005 Apr 8;7(1):E61-77. Review.PMID: 16146351 PubMed - indexed for MEDLINEFree PMC ArticleFree textRelated citations27當(dāng)前最佳證據(jù)表1-3評(píng)價(jià)的

13、基礎(chǔ)上應(yīng)用28研究設(shè)計(jì)（依證據(jù)級(jí)別）治療診斷病因預(yù)后系統(tǒng)評(píng)價(jià)隨機(jī)對(duì)照試驗(yàn)非隨機(jī)的對(duì)照試驗(yàn)隊(duì)列研究（觀察）無(wú)對(duì)照的病例系列個(gè)案報(bào)告系統(tǒng)評(píng)價(jià)隊(duì)列研究系統(tǒng)評(píng)價(jià)隨機(jī)對(duì)照試驗(yàn)隊(duì)列研究病例-對(duì)照研究系統(tǒng)評(píng)價(jià)隊(duì)列研究病例-對(duì)照研究29系統(tǒng)評(píng)價(jià) Systematic Review選題：要回答的臨床問(wèn)題擬定納入標(biāo)準(zhǔn)和排除標(biāo)準(zhǔn)檢索：檢索策略和數(shù)據(jù)庫(kù)確定納入的試驗(yàn)評(píng)價(jià)納入試驗(yàn)的方法學(xué)質(zhì)量抽取試驗(yàn)的結(jié)果數(shù)據(jù)匯總分析討論總結(jié)： 回答問(wèn)題查找證據(jù) 分析證據(jù) 最高級(jí)別的證據(jù)30/3132常用統(tǒng)計(jì)學(xué)指標(biāo)二分變量（dichotomous)比值比（Odds Ratio, OR）危險(xiǎn)比（Risk Ratio, RR），或稱相對(duì)危險(xiǎn)

14、度（Relative Risk, RR）危險(xiǎn)差（Risk Reduction, RD），或稱絕對(duì)危險(xiǎn)減少（Absolute Risk Reduction, ARR）需要治療的病人數(shù)（Number Needed to Treat, NNT）連續(xù)變量（Continuous）均數(shù)差（Mean Difference, MD）33Risk=發(fā)生某事件的人數(shù)觀察的總?cè)藬?shù) -實(shí)際上指某事件的發(fā)生率odds=發(fā)生某事件的人數(shù)未發(fā)生某事件人數(shù)比如24人沿一斜坡滑雪，6人滑倒Risk=624=0.25Odd=6(24-6)=0.33在事件發(fā)生率很低情況下，risk和odds接近。 Risk and Odds34O

15、RRR干預(yù)組odds / 對(duì)照組odds 干預(yù)組risk / 對(duì)照組risk=1，干預(yù)組和對(duì)照組的事件發(fā)生頻率一致1，干預(yù)使事件發(fā)生頻率增高1，干預(yù)使事件發(fā)生頻率降低離1越近，效應(yīng)值越小，離1越遠(yuǎn)，效應(yīng)值越大。95% CI 如果包括1，說(shuō)明差異無(wú)統(tǒng)計(jì)學(xué)意義（P0.05），否則差異有統(tǒng)計(jì)學(xué)意義（P0.05）。 OR和RR 反映治療組和對(duì)照組事件發(fā)生頻率對(duì)比35可信區(qū)間 （Confidence Interval，CI） 一個(gè)試驗(yàn)的結(jié)果效應(yīng)值點(diǎn)估計(jì)同一試驗(yàn)重復(fù)多次，結(jié)果不一致效應(yīng)值范圍可信區(qū)間是指真實(shí)的效應(yīng)值所分布的范圍，范圍越小，說(shuō)明對(duì)真實(shí)效應(yīng)值的估計(jì)越精確 95% CI 指對(duì)每一項(xiàng)研究估計(jì)的效應(yīng)

16、值的分布范圍，人們有95%的把握認(rèn)為效應(yīng)真值是在此區(qū)間之內(nèi) 如果OR或RR的95% CI 不包括1，說(shuō)明干預(yù)組和對(duì)照組比較，危險(xiǎn)（或比值）的差異有統(tǒng)計(jì)學(xué)意義，P0.05。否則無(wú)統(tǒng)計(jì)學(xué)意義，P0.05。 36雙膦酸鹽的療效比較（vs安慰劑）觀察項(xiàng)目 （研究數(shù)目）雙膦酸鹽對(duì)比安慰劑 (病例數(shù)) Odds Ratio（95可信區(qū)間）p值評(píng)估病例數(shù)結(jié)果Mortality (10) 1,079 vs 1,048 557 vs 549 0.99 (0.88 to 1.12) 0.9Vertebral fractures (7) 575 vs 541 141 vs 188 0.59 (0.45 to 0.7

17、8) 0.0001Non-vertebral fractures (6) 708 vs 681 102 vs 93 1.05 (0.77 to 1.44) 0.7Hypercalcemia (8) 1,044 vs 1,002 84 vs 101 0.76 (0.56 to 1.03) 0.07Pain (8) 657 vs 624 276 vs 318 0.59 (0.46 to 0.76) 0.00005Gastrointestinal symptomsd (6) 853 vs 836 110 vs 86 1.28 (0.95 to 1.74) 0.11373839Odds ratio (

18、OR) for vascular events40疾病疾病暴露41RD干預(yù)組risk 對(duì)照組risk =0，干預(yù)組和對(duì)照組的事件發(fā)生頻率一致0，干預(yù)使事件發(fā)生頻率增高0，干預(yù)使事件發(fā)生頻率降低離0越近，效應(yīng)值越小，離0越遠(yuǎn)，效應(yīng)值越大。95% CI 如果包括0，說(shuō)明差異無(wú)統(tǒng)計(jì)學(xué)意義（P0.05），否則差異有統(tǒng)計(jì)學(xué)意義（P0.05）。 RD 反映治療組和對(duì)照組事件發(fā)生頻率差別42伏立康唑與二性霉素治療侵襲性曲霉菌治療成功的危險(xiǎn)差（RD）43NNT 指獲得（或避免）1個(gè)事件需要治療的病人數(shù)。NNT越大，效應(yīng)值越小 例子：6328人用戒煙口香糖，1149人戒煙成功對(duì)照組8380人，893人戒煙成功口

19、香糖組成功率：1149/6328=18.2%對(duì)照組成功率： 893/8389=10.6%RD=18.2%-10.6%=7.6%NNT=1/RD=1 / 7.6%=13.244RR 對(duì)每個(gè)人而言，“臨床療效”是一樣的，但NNT不同。（表4-3）例子: 某種藥物降低心梗死亡率40%，即 RR=0.60如果治療的對(duì)象人群死于心臟事件的年死亡率1/10，NNT=25如果治療的對(duì)象人群死于心臟事件的年死亡率1/100，NNT=250其他例子（表4-345MD干預(yù)組mean 對(duì)照組mean =0，干預(yù)組和對(duì)照組的測(cè)量值一致0，干預(yù)組測(cè)量值高于對(duì)照組 0，干預(yù)組測(cè)量值低于對(duì)照組 離0越近，效應(yīng)值越小，離0越

20、遠(yuǎn)，效應(yīng)值越大。95% CI 如果包括0，說(shuō)明差異無(wú)統(tǒng)計(jì)學(xué)意義（P0.05），否則差異有統(tǒng)計(jì)學(xué)意義（P0.05）。 MD 反映治療組和對(duì)照組測(cè)量值差別4647診斷研究：Sen/Spe/Acc/+PV/-PV診斷試驗(yàn)金標(biāo)準(zhǔn)合計(jì)有病無(wú)病是ABA+B否CDC+D合計(jì)A+CB+DA+B+C+D敏感度：真陽(yáng)性率，A/(A+C)特異度：真陰性率，D/(B+D)準(zhǔn)確度：(A+D)/(A+B+C+D)陽(yáng)性預(yù)測(cè)值：全部陽(yáng)性中真陽(yáng)性的比例， A/(A+B)陰性預(yù)測(cè)值：全部陰性中真陰性的比例， D/(C+D)受患者集中程度影響：大醫(yī)院都是陽(yáng)性的48診斷研究：似然比(LR)診斷試驗(yàn)金標(biāo)準(zhǔn)合計(jì)有病無(wú)病是ABA+B否CD

21、C+D合計(jì)A+CB+DA+B+C+D陽(yáng)性似然比(+LR):真陽(yáng)性率/假陽(yáng)性率,陽(yáng)性時(shí)患病與不患病機(jī)會(huì)之比陰性似然比(-LR):假陰性率/真陰性率,陰性時(shí)患病與不患病機(jī)會(huì)之比49預(yù)后研究發(fā)病率:一定期間人群新發(fā)病例頻率患病率:特定時(shí)間內(nèi)新舊病例數(shù)/同期平均人口中位生存時(shí)間:50%隨訪者死亡所需隨訪時(shí)間生存率:一段時(shí)間后存活的病例數(shù)/總例數(shù)生存曲線:預(yù)后隨時(shí)間變化的情況50循證臨床實(shí)踐的步驟方法 提出臨床問(wèn)題查詢有關(guān)的研究資料（證據(jù)）評(píng)價(jià)證據(jù)的質(zhì)量（可靠性），了解證據(jù)結(jié)果是什么（效應(yīng)值及其精確度）？這些證據(jù)是否適合于您的病人將這些證據(jù)結(jié)合專業(yè)知識(shí)和病人的要求進(jìn)行醫(yī)療決策評(píng)價(jià)在臨床實(shí)踐中應(yīng)用證據(jù)的效

22、果51Essential steps of EBM -David SackettTo convert our information needs into answerable question (that is, to formulate the problem)To track down, with maximum efficiency, the best evidence with which to answer these questions.To appraise the evidence critically (that is, to weigh it up) to assess

23、its validity (closeness to the truth) and usefulness (clinical applicability)To implement the results of this appraisal in our clinical practiceTo evaluate our performance52構(gòu)建臨床問(wèn)題問(wèn)題的類型問(wèn)題的來(lái)源問(wèn)題的構(gòu)建（PICO） 53問(wèn)題類型治療問(wèn)題診斷問(wèn)題病因問(wèn)題預(yù)后問(wèn)題54Aetiology or Causation: What is the risk that this exposure will cause a gi

24、ven disease?什么是風(fēng)險(xiǎn),這些資訊會(huì)將導(dǎo)致一個(gè)給定的疾病嗎?Diagnosis: what test should I use to investigate this patients pathology and how accurate is that test?診斷:我要用什么測(cè)試研究這個(gè)病人的病理學(xué)和精確性如何這個(gè)測(cè)試嗎?Treatment and Prevention: Does this intervention improve this condition in patients such as this one?防治:這是不是干預(yù)患者改善這一狀況,例如有嗎? Progn

25、osis: Given this patients demographic factors, what is the natural history of this condition?預(yù)后:考慮到患者的人口因素,什么是自然歷史的情況嗎?55背景問(wèn)題前景問(wèn)題56診斷下肢深靜脈血栓的方法小腿腫鑒別診斷DVT蜂窩織炎 B超陰性能排除是DVT嗎？D-Dimer陰性能排除DVT嗎？靜脈造影B超D-Dimer57知識(shí)更新的方式基于主題的學(xué)習(xí)系統(tǒng)以教材為中心效率低下缺乏目的性基于問(wèn)題的學(xué)習(xí)零散以學(xué)習(xí)者為中心印象深刻針對(duì)性強(qiáng)58治療問(wèn)題某種疾病“最好”的治療是什么？什么樣的治療措施“起效最快”、“效果最好”

26、、“副作用最少”？ 59PICO形式Patients or PopulationInterventionComparisonOutcomes60PICO的目的搞清楚問(wèn)的問(wèn)題是什么？明白要回答這個(gè)問(wèn)題需要什么樣的信息？可以根據(jù)問(wèn)題選擇合適的檢索詞?？梢詳M訂和修改檢索策略。 61A question about selecting therapy一個(gè)關(guān)于選擇的治療方法In coronary patients who develop heart failure on ibuprofen, would simple drug withdrawal plus diuretics suffice or s

27、hould they be offered an angiotensin-converting enzyme (ACE) inhibitor, in order to alleviate the heart failure?在冠狀動(dòng)脈患者在布洛芬發(fā)展為心衰,簡(jiǎn)單的藥物戒斷加上利尿劑夠或應(yīng)提供了血管緊張素酶(ACE)抑制劑,為了緩解心臟的失敗?PIC O62A question of selecting a diagnosis test選擇的問(wèn)題診斷測(cè)試In patients with heart failure, how accurate is the bed-side test for po

28、st-Valsalva systolic overshoot in predicting left ventricular function?P: Participants(對(duì)象), pts with heart failureI: Indicator(指標(biāo)), C: ComparisonO: Outcome63A question about predicting prognosis關(guān)于判斷預(yù)后的問(wèn)題In an elderly woman with coronary heart disease, by how much does the appearance of heart failure

29、 shorten her life? P: ParticipantsI: Prognostic indicatorC: ComparisonO: Outcome64A question of determining etiology確定集合的問(wèn)題In an elderly woman with preexisting coronary heart disease, can ibuprofen precipitate heart failure? P: ParticipantsI: ExposureC: ComparisonO: Outcome65PICO治療急性心梗病人早期他汀治療安慰劑死亡診

30、斷懷疑冠心病人ECT心肌顯影冠脈造影敏感度、特異度病因（不良反應(yīng)）絕經(jīng)期婦女激素替代治療無(wú)激素替代乳腺癌發(fā)生預(yù)后充血性心力衰竭年齡70年齡10或0.1，驗(yàn)前概率到驗(yàn)后概率起決定性變化，基本可以確定和排除診斷。122診斷研究的臨界點(diǎn)取值表6-3受試者工作特性曲線（Receiver operator characteristic curve, ROC)123Sen1-Spe124診斷研究的評(píng)價(jià)結(jié)果是否可靠？結(jié)果是什么？結(jié)果適合于當(dāng)前病人嗎？125Appraisal of Articles about Diagnostic Tests Adapted from Duke University Evi

31、dence-based Medicine Website126Are the results valid?Was there an independent blind comparison with a reference standard? 獨(dú)立盲法Were patients from an appropriate spectrum for whom the test will be applied in clinical practice?疾病譜 Was the reference standard performed in all cases or did the results of

32、the test being evaluated influence the decision to perform the reference standard? 是否全部病例均接受標(biāo)準(zhǔn)檢查Were methods of the test described in sufficient detail to permit replication? 采用的篩查方法是否詳述127What are the results?What is the strength of the outcome? What were the likelihood ratios? Sen (95% CI)Spe (95%

33、 CI)Positive predictive value (95% CI)Negative predictive value (95% CI)Likelihood ratios128Were the results help me in caring my patients?Will the reproducibility of the test result and its interpretation be satisfactory in my setting? 129預(yù)后研究預(yù)后：關(guān)于預(yù)測(cè)疾病可能結(jié)局及其發(fā)生的可能性大小的估計(jì)。預(yù)后因素：與病因因素不同。預(yù)后因素包括預(yù)后研究的設(shè)計(jì)類型（

34、表7-3）預(yù)后研究的評(píng)價(jià)（表7-1）130Prognostic factorsPrognostic factors are characteristics of a patient that may be used to more accurately predict that patients outcome. These characteristics may be demographic（人口統(tǒng)計(jì)學(xué)）, disease specific（疾病相關(guān)性）, or co-morbid（并發(fā)?。? Unlike aetiological or causative factors, prognos

35、tic factors need not be causal, just strongly associated with the outcome and thus, an accurate predictor of how an individual may progress through a disease process. 131Prognostic factors are investigated and established via the following types of study:Cohort studies: Participants are selected acc

36、ording to exposure to prognostic factor. They can be studied either prospectively or retrospectively. Case-control studies: Participants are selected by outcome, with the investigator looking back to see if they have the prognostic factor, as compared to a similar control group without the outcome.

37、132Cohort Studies are preferable, as case-control studies rely on the patient or a family member reporting exposure to a prognostic factor. Case-control studies are, therefore, subject to recall bias (回憶偏倚). In addition, case-control studies have potential for bias when selecting cases and controls;

38、 i.e., in defining whether or not the participant has the outcome.133Worksheet for Appraisal of an Article about PrognosisAdapted from Sackett June 1996134Are the results of this prognosis study valid?Was a defined, representative sample of patients assembled at a common (usually early) point in the

39、 course of their disease? 隊(duì)列有代表性，起點(diǎn)一致Was patient follow-up sufficiently long and complete? 隨訪夠長(zhǎng)，完整Were objective outcome criteria applied in a blind fashion? 結(jié)局的判定是客觀的，采用盲法If subgroups with different prognoses are identified, was there adjustment for important prognostic factors? 如果分組分析，對(duì)重要的預(yù)后因素要進(jìn)行校

40、正135Are the results of this prognosis study important?How likely are the outcomes over time? How precise are the prognostic estimates?136137Kaplan-Meier analysis measures the ratio of surviving subjects (or those without an event) divided by the total number of subjects at risk for the event. 138系統(tǒng)評(píng)

41、價(jià)Systematic review概念：系統(tǒng)評(píng)價(jià)，meta分析 意義系統(tǒng)評(píng)價(jià)的評(píng)價(jià)（表8-1）139系統(tǒng)評(píng)價(jià)（systematic review） 指針對(duì)某一具體臨床問(wèn)題，系統(tǒng)、全面地收集所有臨床研究，采用臨床流行病學(xué)嚴(yán)格評(píng)價(jià)文獻(xiàn)的原則和方法，篩選出符合質(zhì)量標(biāo)準(zhǔn)的文獻(xiàn)，進(jìn)行定性或定量合成(Meta-分析），得出綜合可靠的結(jié)論。其與一般意義的綜述不同。一般意義的綜述可能只是就關(guān)于一個(gè)問(wèn)題的數(shù)篇文獻(xiàn)進(jìn)行概括性的描述，可能帶有主觀性，以支持綜述作者的觀點(diǎn)。 140Meta-分析 是1976年由Glass提出來(lái)的，通過(guò)綜合多個(gè)目的相同且相互獨(dú)立的研究結(jié)果，以提供量化結(jié)果來(lái)回答根據(jù)臨床情況提出的研究問(wèn)

42、題，這是目前進(jìn)行系統(tǒng)評(píng)價(jià)的一種研究手段和方法。系統(tǒng)評(píng)價(jià)有時(shí)與Meta分析這個(gè)名詞通用，但嚴(yán)格意義上說(shuō)二者是不同的。系統(tǒng)評(píng)價(jià)可以用、也可以不用meta分析，標(biāo)以meta分析的文獻(xiàn)可能做過(guò)系統(tǒng)全面的檢索（此時(shí)可以說(shuō)是系統(tǒng)評(píng)價(jià)），也可能沒(méi)有。141為什么做系統(tǒng)評(píng)價(jià)？作為臨床醫(yī)生的決策參考：解決信息多、時(shí)間少的問(wèn)題；對(duì)RCT的質(zhì)量和結(jié)果有評(píng)價(jià)，結(jié)論可靠性增加；經(jīng)過(guò)meta分析，有綜合的結(jié)果。綜合多個(gè)樣本量較小的試驗(yàn)結(jié)果，使權(quán)重增加。前人工作的總結(jié)為進(jìn)一步的研究提供線索142Human albumin solutionLicensed indications (in UK)hypovolaemic sh

43、ockburnshypoalbuminaemia143Systematic reviewtrials comparing albumin with no albumin for the licensed indicationsfound 32 trials30 trials with data on death144Total (95%CI)Woittiez 1998So 1997Tollofsrud 1995Woods 1993Rackow 1983Grundmann 1982Zetterstrom 1981bZetterstrom 1981aLucas 1980Boutros 1979Vi

44、rgilio 1979Shah 1977Lowe 19770.010.11101001000Albumin better RR Albumin worseAlbumin vs no colloid in hypovolaemiaRR 1.46 (0.97-2.22)Number of deaths- albumin38/256 (14.8%)- control26/278 (9.4%)145Albumin vs. no colloid in burnsGreenhalgh 19950.010.1110100Total (95%CI)Goodwin 1983Jelenko 1978Albumin

45、 better RR Albumin worseRR 2.40 (1.11 to 5.19)Number of deaths- albumin19/81 (23.5%)- control8/82 (9.8%)146Albumin vs. no colloid in hypoalbuminaemia0.1110100Total (95%CI)Rubin 1997Golub 1994 Greenough 1993Kanarek 1992Foley 1990Brown 1988Nilsson 1980Bland 1976Albumin better RR Albumin worseRR 1.69 (

46、1.07 to 2.67)Number of deaths- albumin41/259 (15.8%)- control24/248 (9.7%)147ConclusionThere is no evidence from the available controlled trials that albumin reduces mortality in critically ill patients withhypovolaemic shockburnshypoalbuminaemia148ConclusionThere is a strong suggestion that albumin

47、 increases mortality in critically ill patients withhypovolaemic shockburnshypoalbuminaemia149Appraisal of systematic reviews? A flawed compilation of data is potentially more harmful than a flawed initial study, because people will assume that it is the right answer as it is gold-standard evidence.

48、 150 Important points of a good systematic reviewDid the trial address a clearly focused question with a clearly defined問(wèn)題是否清楚Is all the relevant evidence either in the review or excluded for a valid reason? 包括所有應(yīng)該納入的研究Has there been some assessment of the quality of the included trials/data? 是否做方法學(xué)

49、質(zhì)量評(píng)估How convincing are the results (in terms of size of effect, homogeneity of results, precision of effect)? Do the results apply to my patient population? 結(jié)果可靠嗎結(jié)果是什么能用于當(dāng)前病人嗎151152153臨床指南（clinical guideline） 最好的研究證據(jù)付諸臨床實(shí)踐的橋梁 由官方機(jī)構(gòu)或?qū)W術(shù)團(tuán)體組織制作針對(duì)某一種常見(jiàn)的疾病或臨床問(wèn)題，通過(guò)全面收集并嚴(yán)格評(píng)估現(xiàn)有各種診療措施的研究證據(jù)，提出指導(dǎo)性的建議 在深入研究證據(jù)的基礎(chǔ)

50、上提出的，因此也稱為循證指南（evidence-based guidelines） 需要各方面的專家，工作量大不做證據(jù)說(shuō)明的指南只能算專家意見(jiàn)，可靠性差154BCSH多發(fā)性骨髓瘤指南 (2005)對(duì)于年齡65歲、體力狀態(tài)和器官功能良好的初診病人，HDT加ASCT應(yīng)該作為一線治療策略的一部分（A級(jí)建議，證據(jù)水平b級(jí)）。對(duì)于體力狀態(tài)良好、 65歲的病人，可以考慮HDT加ASCT（B級(jí)建議，證據(jù)水平a級(jí)）。建議采用苯丙氨酸氮芥單用、不加TBI的預(yù)處理（B級(jí)建議，證據(jù)水平a級(jí)）。通常苯丙氨酸氮芥的劑量為200 mg/m2，年紀(jì)大（超過(guò)65-70歲）的病人和腎功能衰竭的病人要減少劑量。155a從隨機(jī)對(duì)照

51、試驗(yàn)的Meta分析獲得的證據(jù)b從至少一項(xiàng)隨機(jī)對(duì)照試驗(yàn)獲得的證據(jù)a從至少一項(xiàng)設(shè)計(jì)很好的非隨機(jī)對(duì)照試驗(yàn)獲得的證據(jù)，包括期臨床試驗(yàn)和病例-對(duì)照研究b證據(jù)來(lái)自至少一項(xiàng)其他類型設(shè)計(jì)很好的準(zhǔn)試驗(yàn)研究，即研究中沒(méi)有安排好的干預(yù)措施，包括觀察研究證據(jù)來(lái)自很好設(shè)計(jì)的非試驗(yàn)的描述性研究證據(jù)來(lái)自隨機(jī)對(duì)照試驗(yàn)的Meta分析或期臨床試驗(yàn)，但這些證據(jù)只有文摘的形式發(fā)表證據(jù)來(lái)自專家委員會(huì)的報(bào)告或意見(jiàn) 和/或 有名望權(quán)威的臨床經(jīng)驗(yàn)BCSH 指南: 證據(jù)水平 156BCSH 指南: 建議級(jí)別A 級(jí)建議證據(jù)水平a、b建議的依據(jù)是：至少一項(xiàng)針對(duì)特定問(wèn)題所做的隨機(jī)對(duì)照試驗(yàn)，試驗(yàn)的質(zhì)量高，而且結(jié)果一致B 級(jí)建議證據(jù)水平a、b、建議的依

52、據(jù)是：針對(duì)建議主題做的、完成較好的臨床試驗(yàn)，但不是隨機(jī)對(duì)照試驗(yàn)C 級(jí)建議證據(jù)水平證據(jù)來(lái)自專家委員會(huì)的報(bào)告 和/或 有名望權(quán)威的臨床經(jīng)驗(yàn)157系統(tǒng)評(píng)價(jià)的制作過(guò)程選題擬定納入標(biāo)準(zhǔn)和排除標(biāo)準(zhǔn)檢索確定納入的試驗(yàn)評(píng)價(jià)納入試驗(yàn)的方法學(xué)質(zhì)量抽取試驗(yàn)的結(jié)果數(shù)據(jù)匯總分析討論總結(jié) 158159BackgroundAlthough non-steroidal anti-inammatory drugs (NSAIDs) have been widely used for the treatment of pain and fever associated with the common cold, there is

53、 no systematic review to assess the effects of NSAIDs in patients with the common cold.ObjectivesTo determine the effects of NSAIDs versus placebo and other treatments on the signs and symptoms of the common cold. To determine any adverse effects of NSAIDs in patients treated with NSAIDs for the com

54、mon cold.Search strategyWe searched the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2009, issue 1) which includes the Acute Respiratory Infections (ARI) Groups Specialized Register;MEDLINE (January 1966 to March 2009); EMBASE (January 1980 to March 2009); CINAHL (J

55、anuary 1982 to March 2009); ProQuest Digital Dissertations (January 1938 to March 2009); KoreaMed(January 1958 to March 2009) and KMbase (January 1949 to March 2009).Abstract160Selection criteriaRandomized controlled trials (RCTs) studying treatment of the common cold with NSAIDs in adults or childr

56、en. Data collection and analysis Four review authors extracted data (SYK, YSM, YJC, YWH). We subdivided trials into placebo-controlled RCTs and NSAIDs versus NSAIDs RCTs. We extracted and summarized data on global efcacies: analgesic effects such as reduction of headache and myalgia; non-analgesic e

57、ffects such as reduction of nasal symptoms, cough, sputum and sneezing; and side effects.161Main resultsThis review includes nine RCTs, describing 37 comparisons: six were NSAIDs versus placebo, and three were NSAIDs versus NSAIDs. A total of 1064 patients with the common cold were included. In a po

58、oled analysis, NSAIDs did not signicantly reduce the total symptom score, or duration of colds. However, for outcomes related to the analgesic effects of NSAIDs (headache, ear pain, and muscle and joint pain) NSAIDs produced signicant benets, and malaise showed a borderline benet, although throat ir

59、ritation was not improved. Chills showed mixed results. For respiratory symptoms, cough and nasal discharge scores were not improved, but the sneezing score signicantly improved. We found no evidence of increased frequency of adverse effects in the NSAID treatment groups.Authors conclusionsThe autho

60、rs recommend NSAIDs for relieving discomfort or pain caused by the common cold. Further research is needed to investigate the effect of NSAIDs in relieving respiratory symptoms such as cough and nasal discharge.162Sulfasalazine for ankylosing spondylitis J Chen, C LiuCochrane Library 2005, Issue 3Is