Discussion On The Related Factors Influencing The Use Of Evidence In Randomized Controlled Trials Of Traditional Chinese Medicine

ObjectivesCochrane systematic review is a high level evidence source of evidence-based medicine,which has important international influence on guiding medical decisions.By retrieving systematically and assessing strictly randomized controlled trial(RCTs)of traditional Chinese medicine(TCM)in the treatment of type 2 diabetes mellitus(T2DM)published in recent years,this study is aim to explore and analyze factors making the RCTs of TCM for T2DM do not accord with inclusion criteria of the Cochrane systematic review of TCM for T2DM,to provide data support for improving the utilization value of clinical trials of TCM and making the high level of systematic review and evidence-based clinical practice guidelines,and to provide the basis for improving the quality of clinical trials and promoting clinical trials of TCM convert into a valid evidence of evidence-based medicine.Methods1.The use value and influence factor analysis of randomized controlled trials on traditional Chinese medicine in the treatment of type 2 diabetes mellitusWe comprehensively searched all RCTs on TCM for type 2 diabetes in the China National Knowledge Infrastructure(CNKI),Chongqing VIP Chinese science and technology periodical database(VIP),Wanfang Database,Chinese Biomedical Database(SinoMed),Pubmed and Cochrane CENTRAL from their inception to June 2017.After duplicate checking of title,the two researchers examined the titles and abstracts identified potentially eligible trials,and full papers were retrieved and included all RCTs of TCM for type 2 diabetes.According to a Cochrane systematic review of TCM in the treatment of type 2 diabetes,the inclusion criteria whether RCTs conform to the Cochrane systematic review are developed.RCTs were divided into meet or not meet inclusion criteria two groups.Themethodological quality of RCTs was evaluated by the modified Jadad scale.Differences between two groups of RCTs in the literature characteristics such as funds,methodological quality,and so on are compared.Factors making RCTs does not meet to the inclusion criteria of Cochrane systematic review are analyzed,and proportion and years distribution of factors also are analyzed.Finally,whether the proportion of the factors related to literature characteristic is tested.2.Analysis of citations of randomized controlled trials on traditional Chinese medicine for type 2 diabetes mellitusAccording to the two groups of RCTs included in the first part,the citation counts of RCTs were retrieved.Since the same RCTs may be included in multiple databases,after pre-extraction,the RCTs of Chinese shall be checked in the order of CNKI,Wanfang,VIP and SinoMed.The RCTs of English are checked in the order of Google scholar and Web of Science.As the citation counts of RCTs increases over time,the retrieval time is limited to 3 days to reduce the impact of time on citation counts.After retrieving citation counts of all RCTs,the RCTs of Chinese and English with top three citation counts in the two groups were selected.Study design,intervention,control and outcomes and results of these RCTs are extracted,and to retrieve reference these RCTs of literature,namely the source documents.Type of journal and study,and content of the reference of the source item are also extracted.Data were extracted by Excel and analyzed by SPSS 20.0 for descriptive analyses and x2 test,Fisher and non-parameter test were used when necessary.Citation counts distributions of RCTs in two parts are compared.And difference in study characteristics,type of study of source item and referenced content of RCTs with top three citation counts are tested.3.Analysis on rationality of control types in randomized controlled trials for traditional Chinese medicine treating type of 2 diabetes mellitusAccording to the results of the first part,we selected one of the factors that making RCTs in the treatment of type 2 diabetes mellitus not meet to the inclusion criteria of the Cochrane systematic review.To analyze the current situation and existing problems of the control type of RCTs and provide references and suggestions for improving the efficacy evaluation of RCTs.The control information includes type of control,the control measures for the placebo,non-drug intervention or medications.If it is TCM,what type is it?According to control setting principle of clinical epidemiologic,the control is classified and evaluated properly.Results1.The use value and influence factor analysis of randomized controlled trials on traditional Chinese medicine in the treatment of type 2 diabetes mellitusThis update retrieves from January 2015 to June 2017 has retrieved 15741 records and combines the previously retrieved records(from inception to January 2012 and January 2012 to April 2015),a total of 71946 records were retrieved.After screening title and abstract,the majority of studies were excluded with reasons,and full texts of 2828 studies were retrieved.Finally,a total of 2046 RCTs were included,275(13.44%)met inclusion criteriaof Cochrane systematic review and 1771(86.56%)did not meet inclusion criteria.The number of RCTs not meeting inclusion criteria also increased year by year.Although it declined slightly along with the years,RCTs not meeting inclusion criteria still take up a very big part of the proportion(x=0.89,SD:0.08).In RCTs of Chinese(n=2030),the majority of RCTs(1248,61.5%)published in general journal.The proportion of RCTs meeting or not meeting inclusion criteria and both are highest in general journal,while but the proportion of RCTs not meeting inclusion criteria is higher(X=22.2,p<0.001).For funding sources,the proportion of RCTs meeting inclusion criteria is higher RCTs not meeting inclusion criteria(X2=18.86,p<0.001).For sources of trials,the multi-center RCTs with 76.31%(29/38)did not meet inclusion criteria of Cochrane systematic review,and 199(78.66%)of the RCTs in 253(12.46%)RCTs with fund were not eligible for the inclusion criteria.15 RCTs of English were published in journal of Science Citation Index(SCI),which were consistent with inclusion criteria of Cochrane systematic review,and another non-SCI RCT did not conform to the inclusion criteria.RCTs of English have a higher proportion than RCTs of Chinese in number of multi-center trials,sample size estimation,sample size and fund.The major RCTs of Chinese were assessed as low quality,only 92(4.5%)RCTs were high-quality trials.The proportion of high quality RCTs in English is higher than that in Chinese.The high quality RCTs meeting inclusion criteria of Cochrane systematic review were higher than those of RCTs not meeting inclusion criteria(X2=33.83,p<0.001).The reason that the RCTs of English that did not meet inclusion criteria of Cochrane systematic review was that there was not report diagnostic criteria.In 1770 RCTs of Chinese,the most frequent factor was treatment period less than 3 months(n=1137),followed by failing to report inclusion criteria(n=841),ineligible control(n=556)s poor description of interventions(n=520),failing to report clinical outcomes or unusable outcomes(n=349),failing to reported diagnostic criteria(n=289),incorrect random sequence generation method(n=125),failing to report treatment period(n=117),inconsistent co-interventions between two groups(n=30).Except the proportion of RCTs failing to reported diagnostic criteria is increasing with years,other factors all show a downward trend with years.Most RCTs existed two(541,30.5%)or three(530,29.9%)above factors which did not meet inclusion criteria of Cochrane systematic review.In RCTs of Chinese,funding,study quality,and type of journal all influenced the proportions of failing to report diagnostic criteria and inclusion/exclusion criteria,poor description of interventions,failing to report clinical outcomes or outcomes unusable,incorrect random sequence generation method.The proportion of ineligible control is difference in low and high quality trials,but is not affected by funding and type of journal.Conversely,ineligible treatment period is affected by funding and type of journal,not affected by study quality.2.Analysis of citations of randomized controlled trials on traditional Chinese medicine for type 2 diabetes mellitusFor citation counts,Most of RCTs of Chinese have been retrieved in CNKI(n=1828),and all RCTs of English have been retrieved in Google scholar.There was no difference in frequency distribution between the two groups of RCTs(p=0.536).In RCTs of Chinese,citation counts mainly distributed in one to ten times,the number of RCTs that were not referenced at one time showed no difference between the two groups(x2=2.295,p=0.107),but the proportion of RCTs that were not referenced were both larges in two groups(29.23%,76/260;32.94%,583/1770).As citation counts increased,the number of RCTs in both groups gradually decreased.In RCTs with top three citation counts,the citation counts of "eligible RCTs of English"are higher than "eligible RCTs of Chinese",but the total citation counts of "ineligible RCTs of Chinese,are higher than "eligible RCTs of Chinese".The RCTs with top three citation counts in the three groups of RCTs were both placebo controlled and the results were mostly positive.Most of "ineligible RCTs of Chinese,are cited by traditional review,and the proportion that“ineligible RCTs of Chinese,are cited by the systematic reviews is lower than"eligible RCTs of English,and "eligible RCTs of Chinese".The proportion that results of"eligible RCTs of Chinese" were cited is highest,while the proportion that discussion of"ineligible RCTs of Chinese" was cited is higher than the other two groups.3.Analysis on rationality of control types in randomized controlled trials for traditional Chinese medicine treating type of 2 diabetes mellitusControl were classified five types,included hypoglycemic drugs(most are composed of hypoglycemic agents,diet control and exercise),conventional treatments for diabetes,traditional Chinese medicine,placebo-controlled and other,respectively.The proportion of"TCM plus hypoglycemic drugs contrast hypoglycemic drugs" is largest,and most of them are Chinese RCTs,followed by head to head comparison between TCM and hypoglycemic drugs.The proportion that RCTs of Chinese use placebo as a control is low(61/2030,2.86%),while RCTs of English is higher(10/16,62.50%),difference between two groups was statistically significant(continuitycorrection x2=150.46,p<0.001)Hypoglycemic drugs has been the main type of control,and the proportion in total RCTs increasing with years,follow by conventional treatment described by RCTs with the proportion decreasing with years,.TCM as a control continued to decline after 2011(less than 10%).Placebo control is low and has not changed significantly over the years.ConclusionsThe number of RCTs on TCM in the treatment type of 2 diabetes mellitus has increased with years.Although the proportion of RCTs not meeting inclusion criteria of Cochrane systematic review has decreased,the quantity is still large.Ineligible diagnosis and inclusion criteria,interventions,controls,outcomes,treatment duration and comparability between groups are the main factors that reduce the use value of RCTs.Trials with better design and reporting can reduce waste of resources.There was no difference between RCTs meeting and not meeting inclusion criteria of Cochrane systematic review in the average citation counts,but RCTs meeting inclusion criteria cited by systematic review,and the results were cited all higher than RCTs not meeting inclusion criteria,but due to citation counts are affected by many factors,we thought quality of RCTs meeting inclusion criteria may be higher than RCTs not meeting inclusion criteria.Due to most of the TCM lack of high quality evidence at present,it is may unreasonable TCM as a control to prove effect of interventions,and placebo-controlled trials with conventional treatments should be added.