1、,第四节 临床文献评价 真实性 重要性 实用性,临床科研设计书的主要内容1.研究题目：主要研究问题和次要研究问题，说明课题的科 学性与可行性。2.样本含量估计：计算方法和本研究采用样本量的依据。3.研究对象的诊断、纳入、排除标准。4.设计方案、随机方法和研究地点。5.效应指标和测量方法。6.控制偏倚的措施。7.统计分析。8.医德问题。,一、真实性评价1.研究证据是否来源于真正的随机对照试验？2.研究对象是否明确限定？3.防治措施的具体内容是否明确？4.是否观察、报告了全部临床结果？5.报道的结果是否包括了全部纳入的病例？6.统计方法的使用是否合适？是否注意了统计 学意义与临床意义？,（一）研究
2、证据是否来源于真正的随机对照试验？ 随机分组的方法是否交代、是否恰当、是否采用 了隐匿？ 试验组和对照组例数是否相等或接近？ 两组的基线资料和其它非试验因素是否可比？ 分层随机的分层因素是否合适？分层各组的样本 量要仔细计算。,（二）研究对象是否明确限定？ 诊断标准 纳入标准 排除标准,（三）防治措施的具体内容是否明确？ 1.药物：剂量、剂型、给药方法与途径、疗程、 相应的配套治疗。 2.保证防治措施的正确实施。 3.沾染和干扰的预防。 4.依从性的检查。,（四）是否观察、报告了全部临床结果？ 效益和危害-治疗作用与毒副反应 效应指标及其观察时间、空间和测量方式是否合适 是否合适的采用盲法,（
3、五）报道的结果是否包括了全部纳入的病例？ 确保失访（lost follow)率10%。 如失访率 20%，没有多大的临床价值。 失访率在10%-20%之间：采用意愿治疗 分析(intention-to-treat analysis) 试验组失访人数按无效计算 对照组失访人数按有效计算,（六）统计方法的使用是否合适？是否注意了统计学意义与临床意义？ 统计学意义只表明试验组与对照组之间的差异来自防治措施的本身，只能评价这种差异的真实程度，但并不表明疗效差异大小的临床意义。CER ARR NNT的应用。,二、重要性评价真正的阴性结果真正的尚有争议的结果真实有效的结果,（一）效果究竟有多大？明确试验组
4、与对照组事件发生率（如治愈率、死亡率、不良反应等）各有多大，以及组间的差值，对这些差值的临床意义做出评价。,（二）试验效果的准确度如何？常用95%的可信区间表示（95%CI confidence interval)可信区间越小，可信度越靠近真值，反之可信度就要差一些。,三、实用性评价患者的情况是否与我们的相一致？可行性如何？干预措施的“利”与“弊”,（一）试验的证据是否与我们经治的患者情况一致？根据患者的病情、病理损害程度、社会人口特点（人种、性别、年龄等）、亚组的分析特点“对号入座”。,（二）试验结果的可行性如何？技术的可行性病人接受的可行性经济的可行性,（三）试验结果的“利”与“弊”对试验
5、结果施加给病人后可能获得多大程度的效果和引起不良反应的风险度要权衡比较，只有在确定“利”肯定大于“弊”时才能接受、实践这种试验结果。Likelihood of being helped versus harmed LHH,A Multicenter Study of Grepafloxacin (格雷沙星）and Clarithromycin in the Treatment of Patients With Community-Acquired Pneumonia* S. Moola, MBChB; Lars Hagberg, MD, PhD; Gavin A. Churchyard, MB
6、BCh; Joe S. Dylewski, MDCM; Sangeeta Sedani, BSc and Heather Staley, BSc Chest. 1999;116:974-983.,Study objectives: To compare the efficacies of 10-day regimens of grepafloxacin (GFX) (Raxar or Vaxar; Glaxo Wellcome; Greenford, UK), 600 qd, and clarithromycin (CLA) (Klacid, Biaxin, or Klaracid; Abbo
7、tt Laboratories; Chicago, IL), 500 mg bid, in patients with community-acquired pneumonia (CAP), on the basis of clinical response, including radiographic evidence, and bacteriologic efficacy.,Patient Population and Study DesignThis was a phase IIIb, double-blind, randomized, prospective, parallel-gr
8、oup, comparative study (protocol GFXB3003) conducted at 58 centers in 11 countries (Australia, Canada, Czech Republic, Germany, Italy, Israel, New Zealand, Poland, South Africa, Spain, and Sweden). Regulatory approval was obtained where appropriate, and the study was approved by local ethics committ
9、ees. Written, informed consent was obtained for each participant in accordance with national guidelines and the Declaration of Helsinki (Hong Kong amendment, 1989).,Patients with chest radiographs taken within 2 days of the start of study medication confirming pulmonary infiltration or consolidation
10、 （实变）likely to be caused by pneumonia, and patients presenting with one or more of the clinical signs and symptoms consistent with CAPpleuritic chest pain, cough, fever ( 38C), and auscultatory findings such as rales and/or evidence of consolidationwere included in the study.,Sputum production was n
11、ot a requirement for study entry; however, when patients were producing sputum, a sample was collected for culture. Patients could be treated in the community or could be admitted to the hospital, depending on the standard medical practice in different countries.,Patients were excluded if they had n
12、osocomial （院内）pneumonia; required immediate IV antibiotic therapy; had received antibiotic therapy within 3 days before study entry; or had bronchial carcinoma, empyema（脓胸）, lung abscess, uncontrolled asthma, pulmonary tuberculosis, or cystic fibrosis.,As well as other standard exclusion criteria fo
13、r clinical trials, patients with an immunocompromised status, malabsorption （吸收障碍）syndromes, hepatic or renal impairment, and history of seizure（癫痫） disorders were excluded, as were those with known sensitivity to any quinolone or macrolide antibiotic.,Patients were randomized to receive either oral
14、 GFX, 600 mg qd for 10 days (251 patients) or oral CLA, 500 mg bid for 10 days (253 patients). In addition to the active medication, patients received concurrent placebo for study-blinding purposes. Administration of additional antimicrobials was not permitted for the duration of the study, and a re
15、cord was kept of any medication taken concomitantly.,AssessmentsAfter entry into the study and pretreatment assessment, patients were reassessed during treatment (4 to 6 days after treatment initiation), after treatment (1 to 3 days after completion), and at follow-up (28 to 35 days after treatment
16、completion). At each assessment, patients were evaluated for resolution of signs and symptoms of pneumonia. Before treatment and at follow-up (or withdrawal), chest radiography and physical examinations were performed.,At each assessment, sputum samples were collected (if available) for Grams stain,
17、 culture, and susceptibility testing (only samples meeting the recognized criteria of 25 neutrophils and 10 epithelial cells per low-power field were cultured). Blood was collected for culture at the pretreatment assessment, and if the results were positive or if the subject remained febrile, blood
18、was collected for culture at the posttreatment and follow-up assessments.,Primary identification of isolated pathogens was performed using routine laboratory culture methods. Bacterial isolates were tested by disk diffusion for susceptibility to GFX and CLA; in addition, the minimal inhibitory conce
19、ntration was determined. Production of -lactamase was determined by the nitrocefin method where appropriate. For S pneumoniae isolates, susceptibility testing to penicillin was performed.,In addition, serology was performed by a central laboratory on pretreatment and follow-up blood or urine samples
20、 for detection of the atypical respiratory pathogens Mycoplasma pneumoniae, Chlamydia pneumoniae（衣原体）, and L pneumophila. The presence of Chlamydia and Mycoplasma spp was established using indirect fluorescent antibody assay kits for both IgG and IgM (MRL Diagnostics; Cypress, CA and Zeus Scientific
21、; Raritan, NJ). Tests were interpreted from comparisons of pretreatment and follow-up samples according to the instructions of the manufacturer.,At least a fourfold rise in reciprocal (互补）value titer for IgG or IgM from pretreatment to follow-up was required to indicate a positive test result. For M
22、 pneumoniae, an IgG titer of 1:128 was required unless a fourfold increase in IgM was also present, when a titer of 1:64 was acceptable. Legionella spp were detected using radioimmune assays (Bimax; Portland, ME) of urine samples.,C-reactive protein (CRP) determinations were made on blood samples ta
23、ken before and after treatment to evaluate this test as an additional surrogate（替代） marker to indicate the presence of bacterial infection. The number of patients with a CRP value 50 mg/L at the pretreatment and posttreatment assessments was determined and was also analyzed in conjunction with clini
24、cal outcome.,Details of adverse events and any other problems elicited by nonspecific questioning were recorded at each visit.,Efficacy MeasuresThe primary measure of efficacy was the clinical and radiographic response at follow-up; this was assessed for all patients and for patients with documented
25、 infection with either typical or atypical pathogens.,Table 1. Definitions of Clinical Responses Clinical ResponsesDefinitionsCure Improvement or resolution of clinical signs and symptoms after treatment and absence, including radiographic evidence, at follow-upImprovement Improvement but incomplete
26、 resolution of clinical signs and symptoms, including radiographic evidence, at followFailure No improvement during or after treatment or discontinuation of therapy because of a drug-related adverse eventRecurrence Resolution or improvement after treatment with recurrence of clinical symptoms, inclu
27、ding radiographic evidence, at follow-upUnable to be Significant deviations from protocol evaluated,To grade the clinical outcome, the radiographic response was classified at follow-up as resolved (areas of infiltration or consolidation completely clear), improved (areas of infiltration or consolida
28、tion still exist but show evidence of clearing), or unchanged or worse (areas of infiltration or consolidation unchanged or show evidence of spread or increased density).,Table 2. Definitions of Bacteriologic Responses Bacteriologic Responses DefinitionsCure Initial pathogen eradicatedPresumed cure
29、Clinical cure or improvement in the absence of sputum or blood cultureCure with superinfection Eradication of initial pathogen with isolation of new organism(s) associated with clinical symptoms of infectionCure with colonization Eradication of initial pathogen with isolation of new nonpathogenic or
30、ganism(s) not associated with clinical symptoms of infectionRecurrence Initial bacteriologic cure with reisolation of original pathogen at follow-upFailure Initial pathogen not eradicated during treatmentPresumed failure Clinical failure in the absence of sputum or blood cultureFailure with superinf
31、ection Initial pathogen not eradicated plus isolation of new pathogen(s)Failure with resistance Initial pathogen(s) developing resistance during therapyUnable to be evaluated Inability to identify or culture pretreatment pathogens, and any other deviation from protocol,Statistical AnalysisA satisfac
32、tory response rate (clinical cure or improvement) of 80 to 89% at follow-up with a broad-spectrum antibiotic was assumed. Assuming an inability to evaluate rate of 25%, at least 450 patients were required to establish equivalence in both the intent-to-treat (ITT, patients who were randomly assigned
33、and received at least one dose of the study medication) and clinically evaluated populations (eg, patients who were able to be clinically evaluated (CE) in accordance with the study protocol criteria),using a 95% confidence interval (CI) calculated for a two-tailed test of significance at the 15% le
34、vel with 90% power.9 Equivalence was demonstrated if the lower limit of the 95% CI of the difference in response rate among patients receiving GFX minus the response rate to CLA was -15%.,Primary efficacy data were analyzed for the ITT and CE populations. Analyses were also performed on the microbio
35、logic intent-to-treat (MITT; patients in the ITT population who had a pathogen isolated on study entry) and the microbiologically and clinically evaluated (MCE; patients in the MITT population who were able to be clinically evaluated) study populations.,Results:1. Patient Demographics and Dispositio
36、n2. Primary Assessment of Clinical Response at the Follow-Up Visit3. Posttreatment Assessment of Clinical Response4. Assessment of Clinical Response in Patients With Confirmed Infection5. Bacteriologic Response6. Pathogens Isolated and Antibiotic Susceptibility7. Safety,Results Patient Demographics
37、and DispositionPatients were recruited at 58 centers in 11 countries; 43% of the patients were treated in the community setting (general practice or outpatient clinic), and 57% of patients were hospitalized in accordance with the standard medical management of such patients in different countries.,P
38、atient demographics were similar with respect to age, sex, and ethnic origin between the two study groups (Table 3 ). The majority of patients, 73%, were 35 years old, with 23% being 65 years old. Sixty-two percent of patients had a preexisting medical condition on entry into the study, with cardiov
39、ascular (23%) and respiratory (17%) conditions being the most common. The most commonly reported pretreatment symptoms were cough and adventitial sounds, recorded for 95% and 85% of the patients, respectively.,Table 3. Summary of Demographic and Baseline Characteristics*Patient Characteristics GFX,
40、n = 251 CLA, n = 253 Total, n = 504SexFemale 106 (42.2) 98 (38.7) 204 (40.5)Male 145 (57.8) 155 (61.3) 300 (59.5)Age, yr 47.7 17.9 49.2 18.0 48.5 18.0Ethnic origin Asian 4 (1.6) 7 (2.8) 11 (2.2)Black 40 (15.9) 41 (16.2) 81 (16.1)White 198 (78.9) 202 (79.8) 400 (79.4)Other 9 (3.6) 2 (1.2) 12 (2.4)Hei
41、ght, cm 169.8 9.4 170.3 9.6 170.1 9.5 Weight, kg 72.89 16.36 73.45 16.54 73.17 16.43,A total of 504 patients were recruited to the study (ITT population), of whom 251 patients were randomized to receive GFX and 253 to receive CLA. There were 106 withdrawals (21%) from the study (51 GFX, 55 CLA; p =
42、0.78) because of lack of efficacy (4%), adverse events (7%), failure to return for assessment (4%), or other reasons (6%; Fig 1 ). Compliance was good, with 99% of urine test results being positive for the presence of antibiotic and 84% of patients taking their medication in accordance with the prot
43、ocol.,HELP with high resolution image viewing Return to Article Figure 1. Patient flow through trial. *Two patients in the GFX group and four in the CLA group had both typical and atypical pathogens present before treatment. Return to Article,Radiographic findings and the nature and severity of sign
44、s and symptoms were comparable in the two treatment groups. At the pretreatment assessment, 78%, 43%, 48%, and 48% of patients reported moderate or severe cough, dyspnea, pleuritic chest pain, and chills, respectively, compared with 3%, 3%, 1%, and 1% of patients at follow-up. No differences were demonstrated between the two treatment groups. Likewise, adventitial sounds, dullness to percussion, friction rub, arthralgia, and myalgia were markedly reduced at follow-up. Sputum was purulent or mucopurulent in 43% of patients at pretreatment and in only 3% of patients at follow-up,
