""" Chinese Meta-analyses, Especially of Genetic Associations
Another way to look at the increase of systematic reviews and meta-analyses is to examine their geographic provenance to understand which countries are mostly responsible for the rapid growth of this factory.In 2014, of the 9,135 articles classified as “meta-analyses” in PubMed,over a third (n = 3,150, 34%) have author affiliations from China,making China the most prolific producer of English-language, PubMed-indexed meta-analyses. The United States is a remote second with only822 meta-analyses (9%). A dramatic change in the geography of meta-analysis happened in a very short period of time. In 2005, meta-analysesfrom China were rare, and meta-analyses from the United States weremore than 15 times more common (n = 539 from the United States vsn = 33 from China). By 2012 China had surpassed the United States inproduction. Currently, China is publishing 4 times more meta-analysesthan the United States, and the gap continues to widen. This is a changeof epidemic proportions (Figure 2).An empirical evaluation published in 2013 tried to understand whatwas driving this rapid growth.71 The rise of meta-analyses from Chinapertains to all types of meta-analyses, including those of randomized tri-als, epidemiological studies, diagnostic test studies, and any other kind ofdesign. This trend has continued in the time since that empirical evalua-tion; for example, in 2014, among meta-analyses of trials and treatments(those identified with the search string “trial OR randomi OR treat-ment*”), 27% of the total come from China. The strong impetus of Chinato become a major power in biomedical research (and beyond), the in-centives to publish in English-language and PubMed-indexed journals,and the large numbers of emerging Chinese authors have buttressedthis epidemic growth. The share of China in meta-analyses has beengrowing much faster than in other types of publications, for example,trials, epidemiological studies, or bench research.71 Perhaps the reasonsfor this are that meta-analyses can be done with little or no money;they have acquired large impact and importance for biomedicine andhealth care as the top of the pyramid in most hierarchies of evidence they can be published in prestigious journals; and they are often heavilycited.73Interestingly, the most spectacular rise of Chinese meta-analyses hasoccurred in the field of genetics.71 The search string “gene OR geneticOR polymorphism OR genome OR mutation OR haplotype” was usedto identify such articles. By mid-2012, China was already producingabout half of all such meta-analyses.71 In 2014, China published 1,210such genetic meta-analysis articles out of a global total of 1,910, that is,63% of the global production, while the United States published only136 (7%) (Figure 3).The empirical evaluation71 examined more closely these articles fromChina. At face value they looked excellent; that is, their reporting wasdone appropriately, with careful tabulations, and publication venueswere respectable English-language journals. However, almost all of themare likely to have reached misleading conclusions. The reason is that theyused the same recipe (now mostly abandoned) that led to many thousandsof articles with misleading results by American and European teams inthe 1990s and early 2000s: candidate gene studies with single or a fewgenes and variants addressed one at a time, by single teams, with smallsample sizes, and with fragmented reporting of the literature subjectto publication bias. Meta-analyses collating such studies almost alwaysgive nominally statistically significant results at p < 0.05, but thismeans very little based on what is known in the current era of genomics.In fact, almost 99% of these claimed associations were not validatedwhen tested in very large consortia with very large sample sizes and noselective reporting, where the entire genome was assessed. """
Repost this lynk
Chinese Meta-analyses, Especially of Genetic Associations
Another way to look at the increase of systematic reviews and meta-analyses is to examine their geographic provenance to understand which countries are mostly responsible for the rapid growth of this factory.In 2014, of the 9,135 articles classified as “meta-analyses” in PubMed,over a third (n = 3,150, 34%) have author affiliations from China,making China the most prolific producer of English-language, PubMed-indexed meta-analyses. The United States is a remote second with only822 meta-analyses (9%). A dramatic change in the geography of meta-analysis happened in a very short period of time. In 2005, meta-analysesfrom China were rare, and meta-analyses from the United States weremore than 15 times more common (n = 539 from the United States vsn = 33 from China). By 2012 China had surpassed the United States inproduction. Currently, China is publishing 4 times more meta-analysesthan the United States, and the gap continues to widen. This is a changeof epidemic proportions (Figure 2).An empirical evaluation published in 2013 tried to understand whatwas driving this rapid growth.71 The rise of meta-analyses from Chinapertains to all types of meta-analyses, including those of randomized tri-als, epidemiological studies, diagnostic test studies, and any other kind ofdesign. This trend has continued in the time since that empirical evalua-tion; for example, in 2014, among meta-analyses of trials and treatments(those identified with the search string “trial OR randomi OR treat-ment*”), 27% of the total come from China. The strong impetus of Chinato become a major power in biomedical research (and beyond), the in-centives to publish in English-language and PubMed-indexed journals,and the large numbers of emerging Chinese authors have buttressedthis epidemic growth. The share of China in meta-analyses has beengrowing much faster than in other types of publications, for example,trials, epidemiological studies, or bench research.71 Perhaps the reasonsfor this are that meta-analyses can be done with little or no money;they have acquired large impact and importance for biomedicine andhealth care as the top of the pyramid in most hierarchies of evidence they can be published in prestigious journals; and they are often heavilycited.73Interestingly, the most spectacular rise of Chinese meta-analyses hasoccurred in the field of genetics.71 The search string “gene OR geneticOR polymorphism OR genome OR mutation OR haplotype” was usedto identify such articles. By mid-2012, China was already producingabout half of all such meta-analyses.71 In 2014, China published 1,210such genetic meta-analysis articles out of a global total of 1,910, that is,63% of the global production, while the United States published only136 (7%) (Figure 3).The empirical evaluation71 examined more closely these articles fromChina. At face value they looked excellent; that is, their reporting wasdone appropriately, with careful tabulations, and publication venueswere respectable English-language journals. However, almost all of themare likely to have reached misleading conclusions. The reason is that theyused the same recipe (now mostly abandoned) that led to many thousandsof articles with misleading results by American and European teams inthe 1990s and early 2000s: candidate gene studies with single or a fewgenes and variants addressed one at a time, by single teams, with smallsample sizes, and with fragmented reporting of the literature subjectto publication bias. Meta-analyses collating such studies almost alwaysgive nominally statistically significant results at p < 0.05, but thismeans very little based on what is known in the current era of genomics.In fact, almost 99% of these claimed associations were not validatedwhen tested in very large consortia with very large sample sizes and noselective reporting, where the entire genome was assessed.
"""