Analysis of official deceased organ donation data casts doubt on the credibility of China's organ transplant reform

Abstract

Background: Since 2010 the People's Republic of China has been engaged in an effort to reform its system of organ transplantation by developing a voluntary organ donation and allocation infrastructure. This has required a shift in the procurement of organs sourced from China's prison and security apparatus to hospital-based voluntary donors declared dead by neurological and/or circulatory criteria. Chinese officials announced that from January 1, 2015, hospital-based donors would be the sole source of organs. This paper examines the availability, transparency, integrity, and consistency of China's official transplant data.

Methods: Forensic statistical methods were used to examine key deceased organ donation datasets from 2010 to 2018. Two central-level datasets - published by the China Organ Transplant Response System (COTRS) and the Red Cross Society of China - are tested for evidence of manipulation, including conformance to simple mathematical formulae, arbitrary internal ratios, the presence of anomalous data artefacts, and cross-consistency. Provincial-level data in five regions are tested for coherence, consistency, and plausibility, and individual hospital data in those provinces are examined for consistency with provincial-level data.

Results: COTRS data conforms almost precisely to a mathematical formula (which first appeared to be a general quadratic, but with further confirmatory data was discovered to be a simpler one-parameter quadratic) while Central Red Cross data mirrors it, albeit imperfectly. The analysis of both datasets suggests human-directed data manufacture and manipulation. Contradictory, implausible, or anomalous data artefacts were found in five provincial datasets, suggesting that these data may have been manipulated to enforce conformity with central quotas. A number of the distinctive features of China's current organ procurement and allocation system are discussed, including apparent misclassification of nonvoluntary donors as voluntary.

Conclusion: A variety of evidence points to what the authors believe can only be plausibly explained by systematic falsification and manipulation of official organ transplant datasets in China. Some apparently nonvoluntary donors also appear to be misclassified as voluntary. This takes place alongside genuine voluntary organ transplant activity, which is often incentivized by large cash payments. These findings are relevant for international interactions with China's organ transplantation system.

Keywords: Data falsification; Organ donation; Organ transplantation; Organ transplantation in China; Statistical forensics; Transplant ethics.

Fig. 1

COTRS Central Level Annual Figures Presented by Huang Jiefu in February 2017. The line of best fit, added, is a quadratic formula with a growth curve that takes the form y = ax² + bx + c. The coefficients for the three lines and a full mathematical explanation by which the line of best fit was derived can be found in Additional file 1

Fig. 2

Reduction in SSE by Using y = a.x²+ b.x + c Instead of y = a.x². The figure plots a range of values that the 2017 annual deceased donor figure may have taken in the COTRS 2018 dataset, showing that the actual value (5146) was extremely close to the value that would have most supported simplification of the model

Fig. 3

Sensitivity to q When Fitting the Model y = a.x^q. The figure shows a range of values for q in the COTRS 2017 and COTRS 2018 series for deceased donors, showing that the addition of the 2017 value of 5146 moves q, for the optimal y = ax^q, much closer to 2 (from 2.07 to 2.01) and further reduces 1-r²(adjusted)

Fig. 4

Organs Transplanted per Donor During Each Intervening Period in the Central Red Cross Series Since 4/7/2014. Points on graph are values during the prior time interval. The horizontal dotted line indicates transplants/donor equal to 2.75

Fig. 5

Central Red Cross Data Donors Per Day During Intervals Since 3/31/2014. Points on graph are values during the prior time interval

Fig. 6

Central Red Cross Data Cumulative Rates Per Day Since 4/7/2014. Scales set so that lines coincide when cumulative transplants/donor equals 2.75

Fig. 7

Cumulative Total of Central Red Cross Registered Volunteers

Fig. 8

Cumulative Totals of COTRS and Central Red Cross Data