Prevalence of diabetes recorded in mainland China using 2018 diagnostic criteria from the American Diabetes Association: national cross sectional study

Prevalence of diabetes recorded in mainland China using 2018 diagnostic criteria from the American Diabetes Association: national cross sectional study Yongze Li Di Teng Xiaoguang Shi Guijun Qin Yingfen Qin Huibiao Quan Bingyin Shi Hui Sun Jianming Ba Bing Chen Jianling Du Lanjie He Xiaoyang Lai Yanbo Li Haiyi Chi Eryuan Liao Chao Liu Libin Liu Xulei Tang Nanwei Tong Guixia Wang Jin‐an Zhang Youmin Wang Yuanming Xue Li Yan Jing Yang Lihui Yang Yongli Yao Zhen Ye Qiao Zhang Lihui Zhang Jun Zhu Mei Zhu Guang Ning Yiming Mu Jiajun Zhao Weiping Teng Zhongyan Shan

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Prevalence of diabetes recorded in mainland China using 2018 diagnostic criteria from the American Diabetes Association: national cross sectional study 2020 BMJ 10.1136/bmj.m997 li2020-prevalence-of-diabetes-recorded-in li2020-prevalence-of-diabetes-recorded-in.txt, the project's staged copy 61159bafc9a84263db91e278b40753dd59092805 cc-by-nc epidemiology

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them TWO DIAGNOSTIC DEFINITIONS RUN OVER THE SAME 75,880 PEOPLE - the 2018 American Diabetes Association criteria and the World Health Organization criteria, differing by the inclusion of HbA1c, producing 12.8% and 11.2% prevalence from one national survey Quotation withheld because this record is not cleared for redistribution. what the prevalence of diabetes and prediabetes is in mainland China, and how much of the measured rise since 2007 is the disease and how much is the criteria Quotation withheld because this record is not cleared for redistribution. a nationally representative prevalence estimate reported under BOTH criteria from the same participants, with a like-for-like WHO-criteria series back to 2007 and provincial, ethnic and urban-rural breakdowns Quotation withheld because this record is not cleared for redistribution. existence Quotation withheld because this record is not cleared for redistribution. a disease whose presence is decided by cut-offs on continuous blood measurements, so that the count of people who have it moves when the cut-offs or the assays change Quotation withheld because this record is not cleared for redistribution. observational-study observational-study Quotation withheld because this record is not cleared for redistribution. THE SAME PEOPLE, TWO DEFINITIONS, 1.6 PERCENTAGE POINTS APART: `This large national survey indicated that 11.2% OF ADULTS (according to WHO criteria) OR 12.8% OF ADULTS (ADA criteria including the addition of HbA1c) aged 18 and older living in China had diabetes in 2017. Thus the total number of patients with diabetes in mainland China is estimated to be 129.8 MILLION.` One survey, one blood draw, two answers - and on a population this size the gap between them is on the order of eighteen million adults. simonneau2018 moved a threshold and did not estimate who it admitted; this paper simply reports both Quotation withheld because this record is not cleared for redistribution. AND THE PAPER SAYS THE TREND IS PARTLY THE DEFINITION, IN ITS INTRODUCTION: `The increase in prevalence could also be partially attributable to the new diagnostic criteria, with reduced cut-off points for blood glucose and the addition of glycated haemoglobin (HbA1c) ... The main difference between them is the inclusion of HbA1c in the ADA criteria.` The rise in a disease is stated as partly an artefact of how the disease is defined, by the people measuring the rise Quotation withheld because this record is not cleared for redistribution. THE HEADLINE HISTORICAL SERIES CROSSES TWO RULERS MID-WAY: `Quotation withheld` A sixteenfold rise reported across a change of diagnostic criteria, with both criteria named in the same sentence. The paper's own caveat about differing sampling and screening methods follows immediately and is OUTSIDE THIS SPAN - it is separated from this clause by a run of seven citation numbers in the staged text. Quotation withheld because this record is not cleared for redistribution. AND THE LIKE-FOR-LIKE SERIES IS PRINTED SEPARATELY, WHICH IS THE REPAIR: `The weighted prevalence of total diabetes diagnosed by the WHO criteria has been increasing from 9.7% in 2007 and 2010, to 10.4% in 2013, and to 11.2% in this study in 2017 ... The prevalence of prediabetes was 35.2%, which was similar to the 35.7% reported in 2013 with the same ADA diagnostic criteria.` Every comparison in the discussion names which criteria it holds fixed - `under the same WHO diagnostic criteria`, `with the same ADA diagnostic criteria`. That phrase is the whole methodological contribution Quotation withheld because this record is not cleared for redistribution. AND THE SURVEY DEFINITION IS NOT THE CLINICAL ONE, WITH THE DIRECTION OF THE ERROR NAMED: `we did not carry out repeat testing in people with abnormal glucose values over time. Quotation withheld` donnelly2019 says a research definition must not guide care; this says the research definition and the clinical one give different answers, and which way Quotation withheld because this record is not cleared for redistribution. AND A THRESHOLD IS FLAGGED AS IMPORTED FROM ANOTHER POPULATION: `using levels of 6.5% for HbA1c to diagnose diabetes could be controversial since that level was set for the US population. Although geographic disparities might exist in HbA1c LEVELS, A consistent threshold was preferred across the study to enable comparison.` The choice is comparability over local validity, and it is stated as a choice - which is the same trade parks2021 makes for species and polman2011 refuses for CSF status Quotation withheld because this record is not cleared for redistribution. AND THE GEOGRAPHIC SPREAD DWARFS THE DEFINITIONAL GAP: `Quotation withheld For prediabetes, Yunnan province had the highest prevalence, being three times higher than the lowest prevalence in Anhui province.` 6.2% to 19.9% across provinces against 11.2% versus 12.8% across definitions. Both are real, and only one of them is usually reported as uncertainty Quotation withheld because this record is not cleared for redistribution. AND THE ETHNIC DIFFERENCES ARE MEASURED AGAINST ONE REFERENCE GROUP: `the prevalence of total diabetes and prediabetes in four ethnic groups was significantly different from that in the Han ethnic group. Quotation withheld.` Han 12.8%, Hui 6.3% - a twofold range inside one country, under one definition, in one survey Quotation withheld because this record is not cleared for redistribution. AND A HISTORICAL EXPOSURE IS OFFERED FOR THE AGE GRADIENT: `Quotation withheld Given that the Chinese famine of 1959-61 was among the most severe on record, the affected population comprises those with a current age of 50 and older, who showed a higher prevalence.` The paper adds `Whether or not prenatal exposure to famine affects diabetes remains controversial` Quotation withheld because this record is not cleared for redistribution. AND THE RURAL RISE IS TWO AND A HALF TIMES THE URBAN ONE, UNDER ONE DEFINITION: `Quotation withheld` `Under the same criteria` is what makes the comparison mean anything, and the paper says it every time Quotation withheld because this record is not cleared for redistribution. AND AWARENESS AND PREVALENCE MOVE IN OPPOSITE DIRECTIONS INSIDE ONE TOTAL: `Although a continuing increase in self-reported diabetes was found, a decrease in the prevalence of newly diagnosed diabetes was noted under the same WHO diagnostic criteria ..., suggesting better awareness of diabetes.` The same total can rise while its two components move apart, and only the split shows that the health system changed rather than the disease Quotation withheld because this record is not cleared for redistribution. ONE READING IS TREATED AS THE DISEASE: `we did not carry out repeat testing in people with abnormal glucose values over time. Quotation withheld` The `what is already known` box says the same of the whole literature: `Such surveys generally use a single reading ..., rather than repeat readings which are typical in clinical practice`. Every national prevalence figure for this disease is computed under a definition nobody uses on a patient Quotation withheld because this record is not cleared for redistribution. AND THE SAME IS TRUE OF THE WHOLE SURVEY LITERATURE, IN THE PAPER'S OWN `What is already known` BOX: `Such surveys generally use A single reading of glucose or HbA1c, rather than repeat readings which are typical in clinical practice, where clinical diagnosis is based also on patient symptoms.` Every national prevalence figure for this disease is computed under a definition nobody uses on a patient, and the paper states it as background rather than as its own limitation Quotation withheld because this record is not cleared for redistribution. AND THE HbA1c CUT-OFF WAS SET ELSEWHERE: `using levels of 6.5% for HbA1c to diagnose diabetes could be controversial since that level was set for the US population.` HbA1c is the ONLY difference between the two criteria, so the entire 1.6-point gap rests on a threshold imported from another population Quotation withheld because this record is not cleared for redistribution. AND A SIXTH OF THE COUNTRY IS OUTSIDE THE SAMPLE, IN A DIRECTION THE PAPER NAMES: `non-residents, such as internal migrant workers, who account for more than A SIXTH OF THE NATION'S TOTAL POPULATION and who were more likely to have A LOWER PREVALENCE of diabetes, were not included in the study owing to the study design. These limitations could lead to an overestimation of diabetes in the population.` Both named biases push the same way, and the paper adds them up out loud Quotation withheld because this record is not cleared for redistribution. AND THE COMPARABILITY THAT MAKES THE SERIES WORK IS ITSELF A DESIGN CHOICE: `a similar study design and statistical methods to those used previously make our findings more comparable with historical data. Although lacking a high quality national cohort to estimate the incidence of diabetes, our findings could add value to the assessment of diabetes epidemics.` A series of cross-sections is not an incidence, and the trend is a difference of prevalences rather than a measured flow Quotation withheld because this record is not cleared for redistribution. computing and reporting the prevalence under BOTH the ADA and WHO criteria from the same participants eliminates: a reader having to guess how much of a national figure is the definition · outcome: held Quotation withheld because this record is not cleared for redistribution. attaching 'under the same WHO diagnostic criteria' or 'with the same ADA diagnostic criteria' to every comparison with an earlier survey eliminates: a trend that is really a change of definition · outcome: held Quotation withheld because this record is not cleared for redistribution. keeping one HbA1c threshold across all provinces and ethnicities although it was derived in a different population eliminates: internal incomparability between the study's own subgroups · outcome: partial Quotation withheld because this record is not cleared for redistribution. naming both known selection biases and their shared direction eliminates: an unsigned uncertainty - both the single reading and the excluded migrant workers push the estimate up · outcome: held Quotation withheld because this record is not cleared for redistribution. The American Diabetes Association (ADA) criteria The American Diabetes Association (ADA) criteria Quotation withheld because this record is not cleared for redistribution. World Health Organization (WHO) criteria World Health Organization (WHO) criteria Quotation withheld because this record is not cleared for redistribution. The increase in prevalence The increase in prevalence Quotation withheld because this record is not cleared for redistribution. the new diagnostic criteria the new diagnostic criteria Quotation withheld because this record is not cleared for redistribution. one reading one reading Quotation withheld because this record is not cleared for redistribution. diabetes diabetes Quotation withheld because this record is not cleared for redistribution. Inner Mongolia Inner Mongolia Quotation withheld because this record is not cleared for redistribution. Guizhou province Guizhou province Quotation withheld because this record is not cleared for redistribution. the Tibetan ethnic group the Tibetan ethnic group Quotation withheld because this record is not cleared for redistribution. the Han ethnic group the Han ethnic group Quotation withheld because this record is not cleared for redistribution. prenatal exposure to famine prenatal exposure to famine Quotation withheld because this record is not cleared for redistribution. the risk of hyperglycaemia the risk of hyperglycaemia Quotation withheld because this record is not cleared for redistribution. urban areas urban areas Quotation withheld because this record is not cleared for redistribution. rural areas rural areas Quotation withheld because this record is not cleared for redistribution. `total diabetes`, ADA 2018 - self-reported diagnosis plus newly diagnosed by fasting glucose >=7.0 mmol/L, two-hour glucose >=11.1 mmol/L, **or HbA1c >=6.5%** Quotation withheld because this record is not cleared for redistribution. `total diabetes`, WHO - the same, WITHOUT the HbA1c route. `The main difference between them is the inclusion of HbA1c in the ADA criteria` Quotation withheld because this record is not cleared for redistribution. `newly diagnosed diabetes` (survey sense) - ONE reading, not the repeat testing of clinical practice: `We categorised one reading as being synonymous with diabetes` Quotation withheld because this record is not cleared for redistribution. `prediabetes`, ADA - impaired fasting glucose OR impaired glucose tolerance OR HbA1c 5.7-6.4%, giving 35.2% of Chinese adults Quotation withheld because this record is not cleared for redistribution. extends → the 2007, 2010 and 2013 Chinese national diabetes surveys, whose WHO-criteria series this study continues with a like-for-like estimate Quotation withheld because this record is not cleared for redistribution. critiques → the historical series that runs from 0.67% in 1980 to 10.9% in 2013 across different criteria and different sampling methods Quotation withheld because this record is not cleared for redistribution. critiques → the survey convention of a single reading, which 'may overestimate the prevalence, compared with clinical practice' and is general to the literature Quotation withheld because this record is not cleared for redistribution. extends → the International Diabetes Federation's atlas estimate of 113.9 million adults with diabetes in China, against which this study's 129.8 million is set Quotation withheld because this record is not cleared for redistribution. The American Diabetes Association (ADA) criteria —distinct-from→ World Health Organization (WHO) criteria The American Diabetes Association (ADA) criteria World Health Organization (WHO) criteria the relation itself is separately evidenced · strength of the relation: 0.95 · note: 'The American Diabetes Association (ADA) criteria…' The cleanest natural experiment on a definition in the corpus: same people, same day, same laboratory, two answers - and the paper reports both rather than choosing · the relation stated: the two differ by one route to diagnosis - HbA1c - and that single difference moves the national prevalence from 11.2% to 12.8% in the same participants Quotation withheld because this record is not cleared for redistribution. the new diagnostic criteria —causes→ The increase in prevalence the new diagnostic criteria The increase in prevalence the relation itself is separately evidenced · strength of the relation: 0.9 · note: 'The increase in prevalence…' A causal claim about an epidemiological trend whose cause is a committee. 'Partially' is unquantified here and quantified later by the like-for-like WHO series · the relation stated: part of the measured rise in diabetes is attributed to the criteria themselves - lower cut-offs and the addition of HbA1c - rather than to more disease Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says the increase COULD ALSO BE PARTIALLY attributable to the new criteria; hedged. Quotation withheld because this record is not cleared for redistribution. one reading —same-as→ diabetes one reading diabetes the relation itself is separately evidenced · strength of the relation: 0.85 · note: 'WE CATEGORISED…' An equivalence asserted for operational reasons and disavowed in the same breath - tagged because the paper states it as its own procedure, with the disavowal in the note. The 'what is already known' box generalises it to the whole survey literature · the relation stated: the survey treats a single abnormal measurement as the disease, and says in the same sentence that this overestimates prevalence against clinical practice Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the paper CATEGORISED one reading as diabetes; operationalised-as would fit better than same-as. Quotation withheld because this record is not cleared for redistribution. Inner Mongolia —contrasted-with→ Guizhou province Inner Mongolia Guizhou province the relation itself is separately evidenced · strength of the relation: 0.95 · note: 'In Inner Mongolia…' Worth holding beside the definitional gap: the national figure of 12.8%, which the paper gives elsewhere, describes no province particularly well, and the choice of criteria moves it by about a tenth of the internal range · the relation stated: prevalence runs from 6.2% to 19.9% across provinces under one definition - a threefold spread against the 1.6-point spread between definitions Quotation withheld because this record is not cleared for redistribution. the Tibetan ethnic group —distinct-from→ the Han ethnic group the Tibetan ethnic group the Han ethnic group the relation itself is separately evidenced · strength of the relation: 0.9 · note: 'Compared with the Han…' Four candidate explanations and no way to separate them in a cross-section, which the paper does not claim to do · the relation stated: Tibetan participants had significantly lower diabetes prevalence than the Han reference group, consistent with an earlier survey Quotation withheld because this record is not cleared for redistribution. prenatal exposure to famine —causes→ the risk of hyperglycaemia prenatal exposure to famine the risk of hyperglycaemia the relation itself is separately evidenced · strength of the relation: 0.85 · note: 'One study showed…' The paper immediately adds 'WHETHER OR NOT…', so this edge carries a lower weight and its own contradiction · the relation stated: prenatal famine exposure is reported to raise hyperglycaemia risk in two generations, which would place the effect in exactly the cohort now aged 50+ Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but attributed: the span reports what 'one study showed'. Quotation withheld because this record is not cleared for redistribution. urban areas —contrasted-with→ rural areas urban areas rural areas the relation itself is separately evidenced · strength of the relation: 0.95 · note: 'the national…' 'Under the same ADA diagnostic criteria' is the clause that makes the comparison legitimate, and this paper attaches it to every historical comparison it makes · the relation stated: since 2013 prevalence rose by 1% in urban and 2.5% in rural China under unchanged criteria, closing a gap that used to run the other way Quotation withheld because this record is not cleared for redistribution. 11.2% of adults had diabetes by WHO criteria and 12.8% by ADA criteria including HbA1c, about 129.8 million people 11.2% of adults had diabetes by WHO criteria and 12.8% by ADA criteria including HbA1c, about 129.8 million people Quotation withheld because this record is not cleared for redistribution. AND THE TREND IS PARTLY THE DEFINITION, SAID BY THE PEOPLE MEASURING IT: `The increase in prevalence could also be partially attributable to the new diagnostic criteria, with reduced cut-off points ... and the addition of HbA1c` AND THE TREND IS PARTLY THE DEFINITION, SAID BY THE PEOPLE MEASURING IT: `The increase in prevalence could also be partially attributable to the new diagnostic criteria, with reduced cut-off points ... and the addition of HbA1c` Quotation withheld because this record is not cleared for redistribution. AND THE HEADLINE SERIES CROSSES TWO RULERS: `increased from 0.67% using the WHO criteria in 1980 to 10.9% using the ADA criteria in 2013` - with `different sampling methods and screening procedures` in one clause of caveat AND THE HEADLINE SERIES CROSSES TWO RULERS: `increased from 0.67% using the WHO criteria in 1980 to 10.9% using the ADA criteria in 2013` - with `different sampling methods and screening procedures` in one clause of caveat Quotation withheld because this record is not cleared for redistribution. AND THE HONEST SERIES IS PRINTED SEPARATELY: `diagnosed by the WHO criteria ... 9.7% in 2007 and 2010, to 10.4% in 2013, and to 11.2% in this study in 2017` AND THE HONEST SERIES IS PRINTED SEPARATELY: `diagnosed by the WHO criteria ... 9.7% in 2007 and 2010, to 10.4% in 2013, and to 11.2% in this study in 2017` Quotation withheld because this record is not cleared for redistribution. AND THE SURVEY DEFINITION IS NOT THE CLINICAL ONE: `We categorised one reading as being synonymous with diabetes, which may overestimate the prevalence, compared with clinical practice` AND THE SURVEY DEFINITION IS NOT THE CLINICAL ONE: `We categorised one reading as being synonymous with diabetes, which may overestimate the prevalence, compared with clinical practice` Quotation withheld because this record is not cleared for redistribution. using 6.5% HbA1c to diagnose diabetes could be controversial because that level was set for the US population using 6.5% HbA1c to diagnose diabetes could be controversial because that level was set for the US population Quotation withheld because this record is not cleared for redistribution. diabetes prevalence in Inner Mongolia was three times that in Guizhou diabetes prevalence in Inner Mongolia was three times that in Guizhou Quotation withheld because this record is not cleared for redistribution. AND THE TWO COMPONENTS OF THE TOTAL MOVE APART: `a continuing increase in self-reported diabetes ... [and] A decrease in the prevalence of newly diagnosed diabetes ... under the same WHO diagnostic criteria` AND THE TWO COMPONENTS OF THE TOTAL MOVE APART: `a continuing increase in self-reported diabetes ... [and] A decrease in the prevalence of newly diagnosed diabetes ... under the same WHO diagnostic criteria` Quotation withheld because this record is not cleared for redistribution. excluding internal migrant workers, over a sixth of the population and likely lower in diabetes, could lead to overestimation excluding internal migrant workers, over a sixth of the population and likely lower in diabetes, could lead to overestimation Quotation withheld because this record is not cleared for redistribution. identical except that ADA admits `HbA1c >=6.5%` as a route to diagnosis - `The main difference between them is the inclusion of HbA1c in the ADA criteria` total diabetes Quotation withheld because this record is not cleared for redistribution. 35.2% of Chinese adults (33.5% to 37.0%), `similar to the 35.7% reported in 2013 with the same ADA diagnostic criteria` prediabetes Quotation withheld because this record is not cleared for redistribution. 1.6 percentage points, which on China's adult population is on the order of eighteen million people; the paper's own headcount is 129.8 million on the ADA figure prevalence of total diabetes, same survey, two definitions Quotation withheld because this record is not cleared for redistribution. 1.5 points of real rise over a decade, against 1.6 points of definitional difference in a single year the like-for-like WHO series Quotation withheld because this record is not cleared for redistribution. `Although different sampling methods and screening procedures were used between 1980 and 2013, these data show a remarkable increase` the series that crosses criteria Quotation withheld because this record is not cleared for redistribution. and Han 12.8% against Hui 6.3% across five ethnicities provincial range under one definition Quotation withheld because this record is not cleared for redistribution. excluded for `missing information on sex, age, plasma glucose, or HbA1c` response and exclusions Quotation withheld because this record is not cleared for redistribution. most criteria papers change a threshold and describe the change; this one runs both definitions over the same 75,880 people and prints 12.8% and 11.2% side by side. The comparison is internally valid in a way no cross-study comparison can be - same participants, same assays, same day a definitional difference reported as a number instead of a caveat Quotation withheld because this record is not cleared for redistribution. `under the same WHO diagnostic criteria`, `with the same ADA diagnostic criteria`, `using the WHO criteria in 1980 ... using the ADA criteria in 2013`. The phrase appears wherever two numbers are compared, which is the discipline the corpus's other threshold papers mostly lack every historical comparison names the criteria it holds fixed Quotation withheld because this record is not cleared for redistribution. the single reading `may overestimate the prevalence` and the excluded migrant workers were `more likely to have a lower prevalence` - so `These limitations could lead to an overestimation of diabetes in the population`. A signed uncertainty, which is rarer than an unsigned one two biases, one direction Quotation withheld because this record is not cleared for redistribution. the single variable that separates the two definitions, at a threshold `set for the US population` HbA 1c Quotation withheld because this record is not cleared for redistribution. the two variables that move the prevalence further than the choice of criteria does, and that a single national figure averages away province Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.1136/bmj.m997.