糖尿病及其并发症治疗随机对照试验 文献方法学评价

Complications in Chinese Medical Literatures

目的 :评价国内医学文献中有关糖尿病及其并发症 (以下简称糖尿病 )治疗研究随机对照试验(RCT)的质量 ,为改变和提高临床治疗试验的水平提供依据。方法 :选择国内 6种糖尿病核心期刊 :用手工检索调查 1995 - 1999年间发表的糖尿病治疗的RCT数共 30卷 2 5 0期 ,各杂志发表的RCT数分别为 :中华内分泌杂志 18篇 ;中华内科杂志 5篇 ;中国糖尿病杂志 19篇 ;中国新药与临床杂志 4篇 ;中国中西医结合杂志 35篇 ;中国中西医实用临床急救 13篇。发表的RCT年代分布为 :1995 - 1997年 49篇 ,1998- 1999年 45篇。调查内容包括研究对象选择、随机化方法、盲法、组间可比性、样本大小、依从性、统计学处理、疗效评价指标、副反应观察及随访等。结果 :6种杂志 5年发表的糖尿病治疗RCT共计 94篇。调查表明 ,选择研究对象有明确的纳入标准的有 5 2篇占 5 5 3 % ,有排除标准的有 2 2篇占 2 3 4%。仅有 9篇RCT交待了随机分组的方法 ,没有 1篇交待随机方案隐藏 ,盲法试验有 9篇 (9 6 % ) ,其中双盲为 5篇 (5 3% ) ;80 %采用药物对照 ;17篇采用空白对照(18 10 % ) ;2篇采用安慰剂作对照 (2 % )。治疗组平均样本含量为 32 3± 1(11- 6 6 9)例 ;对照组平均样本含量为 2 8 7± 1 0 5 (6 - 70 7)例。 5 7 45 %有组间可比性 。

WT5”BZ]Objective:To assess the quality of Randomized controlled trials(RCTs)relevant to diabetes mellitus and it's complications published in Chinese medical journal to provide scientific basis in the benefit of improvement of clinical works.Methods:Six Chinese medical journals in the years of 1995-1999 were hand-searched and 94 RCTs for Diabetes mellitus and it's complications were identified from 250 issues of 30 volumes.The journals included Chin Journal of Diabetes;Chin J of new drugs and clinical Remedies;Chin J of endocrine and metabolic;Chin J integrated Trad western Med,and Chin J Internal Med on emergency treatment.Results:In the recent 6 years,the numbers of RCTs were increased.Among the 94 trials there were 55 3% with explicit inclusion criteria for subject seletion and 23 4% with exclusion criteria.Only 9 RCTs described the method for randomized allocation and no one trial had randomized concealment.Blindness and placebo control were used in 9 and 2 RCTs respectively.The average sample sizes of intervention group and control group were 32 3±1(11-669 cases)and 28 7±1 05(6-707cases)separately statistical memagement was mainly assessed on the basis through x 2 or t test.The clinical outcome assessment was made by the normalization of biochemical indicates,elimination of symptoms and sighs but none of them used relative risk reduction(RRR),absolute risk reduction(ARR),number needed to treat(NNT)or intention-to-treat analysis to evaluate treatment effects.Only one trial accounted for compliance of patients,and only in 43% trials the side effects or adverse drug reaction was observed,and none of them had follow-up.Conclusions:More and more RCTs for diabetes mellitus and it's complication with small mediate sample size were published in Chinese Medical literatures.But the quality of the RCTs is not high as placebo controlled and blinding were seldom used and adequate randomization and it's concealment were not used correctly and rigorously.However,pseudo and quasi-randomization should be avoided.Therefore,clinicians should be further trained on clinical epidermiology in order to improve the quality of clinical trial,and the nation-wide and world-wide trials with multi-um centers、largescale and long-term collaboration even with abroad should be made. [WT5”HZ]