Plateletpheresis Donation Trend and a Ban on Family/Replacement Donation in China: A Pseudo-Panel Data Approach Analysis


 Background A nationwide ban on family/replacement donation (FRD) went into effect on April 1, 2018 in China. To date, no reports relevant to the trend of plateletpheresis donations before and after a nationwide ban on FRD were found.Methods We used two independent full samples, consisting of 135851 and 82129plateletpheresis donors from Guangzhou and Chengdu between October 2012 and September 2019, respectively. A pseudo-panel dataset approach was applied by grouping three time-invariant covariates – gender, blood donation history, and birth year across 14 cross-sections (a 6-month interval each)to form a total of 24 cohort groups (14×24=336 cohorts, i.e., cells) with each having common covariates. The outcome was average apheresis platelet units per donor in each cell. We performed a two-piecewise linear mixed model with the cross-section (i.e., time) just right before the ban as a time breakpoint(i.e., 11th cross-section) to examine the trend of outcome with the adjustment of three time-invariant covariates. We removed the family/replacement donations in each of the first 11 cross-sections to detect its possible influence on the trend.Results The final model for the samples from Guangzhou presented a two-piecewise linear trend of the outcome over time with a horizontal line to the left of the breakpoint (βtimeBefore11=0.0111,p=0.0976) and a significantly positive linear trend to the right (βtimeAfter11=0.0404,p<0.0001). The male donors and the donors with plateletpheresis donation history had an increased baseline outcome and a significant outcome change over time after the ban. Such a two-piecewise linear trend pattern can be replicated using the samples from Chengdu with some minor variations. Removing the FRD before the ban can change the pattern.Conclusion The significant increase of the average apheresis platelet units per donor over time after the FRD ban may be related to the implement of the FRD ban and the improved donation behavior of male donors and/or donors with platelet donation history after the ban. Our findings may potentially motivate the policymakers in other countries where the FRD for plateletpheresis donation is still legitimate to phase out their FRD strategy and ultimately achieve 100% voluntary plateletpheresis donation.

the pattern.
Conclusion The signi cant increase of the average apheresis platelet units per donor over time after the FRD ban may be related to the implement of the FRD ban and the improved donation behavior of male donors and/or donors with platelet donation history after the ban. Our ndings may potentially motivate the policymakers in other countries where the FRD for plateletpheresis donation is still legitimate to phase out their FRD strategy and ultimately achieve 100% voluntary plateletpheresis donation.

Contributions To The Literature
It was thought that banning the family/replacement donation (FRD) could cause severe blood shortage in countries with limited resources. However, a phenomenon of "fewer donors, more blood" in plateletpheresis donation practice since the implementation of the FRD ban in China has broken this belief.
To date, no evidence-based research relevant to the FRD ban has been found. It is necessary to evaluate the in uence of the FRD ban on plateletpheresis donation in China.
Ours ndings lled the literature gap and motivated the policymakers to change the FRD policy for plateletpheresis donation in those countries where the FRD is still legitimate.

Background
The adequacy and safety of blood supply is a major public health challenge in the world [1,2]. There is an ongoing debate over the family/replacement donation (FRD) policy in uencing both shortage and safety of blood supply [3][4][5]. FRD, also called mutual donation, occurs when family members are required to donate blood to replace each unit used by their friend or relative [6] Currently, the FRD is legitimate and considered to be indispensable to the transfusion services in many countries with the limited resources [7][8][9]. Although this type of blood donation may provide short-term solutions for dealing with the shortage of blood supply [10], it increases public distrust in voluntary blood donation and affects the quality and safety of donated blood [11]. Phasing out the FRD is one of the targets in a global framework for action to achieve 100% voluntary blood donation developed by the World Health Organization and the International Federation of Red Cross and Red Crescent Societies [9]. A nationwide ban on the FRD went into effect on April 1, 2018 in China. With the "more donors, more blood " belief, it was believed that banning the FRD would cause serious consequences related to the shortage of blood supply [8,12]. However, a phenomenon of "fewer donors, more blood" for the plateletpheresis donation has been emerging in the eld since the implement of the FRD ban in China. To date, no reports relevant to the trend of plateletpheresis donations before and after a nationwide ban on family/replacement donation were found. Therefore, it is necessary to quantify the evidence-based trend of the plateletpheresis donation and the potential contributing factors to this trend before and after the FRD ban in China.
In present study, we examined a model-based trend of the apheresis platelet units donated before and after the FRD ban using two independent full samples in China with a pseudo-panel data approach followed by a piecewise linear mixed model.

Study population and study design
Based on the availability of the data, we used two independent full samples between October 1, 2012 -September 30, 2019 from Guangzhou Blood Center, the second largest blood center in China and Chengdu Blood Center, the largest blood centers in Western China, which were named as "discovery GZ set" and "replicate CD set" at the individual level, respectively. The discovery and replicate sets consisted of 135851 and 82129 plateletpheresis donors, respectively. We enrolled the donors with an age of 18 ~ 60 years old, weight ≥ 50 kg for male and ≥ 45 kg for female, systolic blood pressure 90 ~ 140 mmHg and diastolic blood pressure 60 ~ 90 mmHg, pulse 60 ~ 100 beats/min, normal body temperature, platelet counts before donation 150 ~ 450 × 109/L, and no any other health conditions according to the China national standard for the eligible donor selection criteria (GB 18467 − 2011). All data used in the present study were de-identi ed. We chose 5.5 years and 1.5 years before and after April 1, 2018, respectively, to set our investigation time window with a six-month interval as a cross-section, generating a total of 14 repeated cross-sections with 11 and 3 cross-sections before and after the ban, respectively. Based on the GB 18467 − 2011, a donor may donate up to a total of 24 plateletpheresis collections during a 12-month rolling period. Thus, some plateletpheresis donors may appear more than once across the 14 cross-sections, but not all donors appear in every cross-section. Therefore, there were some different degrees of the correlations of the data across 14 cross-sections, and our datasets presented a pseudo-panel structure [13] .

Measurements of outcome variable and covariates
The records for the total amount of the platelets donated by each donor within each cross-section, grouping covariates -gender, birth year, and blood donation history (see detailed information below), as well as the variable -family/replacement platelet donation (yes vs. no) were extracted from the archived blood donation documents in both blood centers. The outcome variable at the cohort (i.e., cell) level within the pseudo-panel datasets (see detailed information below) was de ned as average platelet units per donor in each cell.

Construction of pseudo-panel datasets
The pseudo-panel data approach is actually a solution to transform the individual-level cross-sectional data into the group-level data (i.e., pseudo-panel data)such that the typical longitudinal models can be applied to e ciently and consistently estimate the change of the interested outcome variable over time. [13] This approach has been increasingly applied to public health [14,15] .
According to the methods described in the literature [16,17], we constructed two pseudo-panel datasets from our discovery and replicate data by grouping three time-invariant variables -gender, birth year, and blood donation history. Other potential covariates such as ethnicity, occupation, and education with large proportions of missing data were excluded from the analyses. Brie y, individual platelet donors were rst classi ed based on gender that had two categoriesmale and female. To balance the size of each cohort (≥ 100, named as large cohort) and the number of large cohorts [16] within our pseudo-panel datasets, we de ned birth year as three categories -"1952-1974", "1975-1984", and "1985-2001". The variable blood donation history was coded as 4 levels -"None", "Whole Blood Donation Only" (abbreviated to "WB"), "Apheresis Platelet Donation Only" (abbreviated to "PLT"), and "Both WB and PLT Donations" (abbreviated to "Both"). The individuals were then further divided by these two variables, generating 2 × 3 × 4 = 24 cohort groups across the 14 cross-sections, i.e., 24 × 14 = 336 cells for each of two pseudo-panel datasets. The generated each cohort group had common gender, birth year, and blood donation history. To reduce the measurement error [17] ,we removed the cells with less than 30 individual donors in each dataset to generate two nal pseudo-panel datasets, named as "overall Guangzhou pseudo-panel set" (abbreviated to "overall GZ set") (n = 330 cells) and "overall Chengdu pseudo-panel set" (abbreviated to "overall CD set") (n = 316 cells), respectively, for further analyses. More detailed information about the summaries of the constructed pseudo-panel datasets is presented in Table 4.
Our study was actually an observational study using pseudo-panel approach to analysis the data [13][14][15]. Thus, we used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement to ensure standardisation and enhance the quality of the reporting [18]. See Additional le 1 for the STROBE statement.

Statistics
For the individual-level data, we used two-tailed independent t-tests (α = 0.05) to compare total number of donors, total number of donations, and total apheresis platelet units (U), average plateletpheresis donations per donor, and average apheresis platelet units (U) per donor between before and after the ban (also called "after-before mean difference"). To compare two after-before mean differences between groups within each covariate, we applied two-tailed Z-test (Z=(mean difference1-mean difference2)/sqrt(se12 + se22), α = 0.05).
For overall GZ pseudo-panel set, we plotted outcome values (i.e., average apheresis platelet units per donor per cell) from each cohort group that had common gender, birth year, and blood donation history versus time (i.e., 14 cross-sections) as well as overall average outcome values from all 24 cohort groups versus time to visualize whether the trend of the outcome over time is linear or non-linear. Given that generalized linear mixed model and mixed-generalized ordered logit model have been successfully applied to analyze the trend of the outcome variable (proportion or probability) over time within a pseudo-panel dataset [14,19], for our pseudo-panel dataset with a continuous outcome variable, we applied a linear mixed model for modelling the linear trend, or modelling the nonlinear trend using a piecewise linear mixed model with the de ned time (i.e., cross-section) breakpoint(s) based on the above-mentioned visualization [20]. We conducted a model selection starting from an unadjusted model (i.e., model 1 -pure time trend model without any covariates) to an adjusted model (i.e., model 2 = model 1 + signi cant covariates), and then to a nal model (i.e., model 3 = model 2 + signi cant interaction terms). To compare two raw regression coe cients within the same nal model, we used one-tailed Z-tests [Z = abs(β2-β1)/sqrt(SE12 + SE22), α = 0.05].
The assumptions of normality and homoscedasticity for piecewise linear mixed effects models were examined by visualizing marginal (for xed effects only) and conditional (for both xed and random effects) Pearson residual plots.
For ve-fold cross-validation of the nal model, the data were randomly split into 5 roughly equal-sized subsets using SAS PROC SURVEYSELECT procedure, the nal model was tted to the 4 subsets of the data using SAS PROC MIXED procedure with the STORE statement. The prediction error -Root Mean Square Error (RMSE) of the tted model to predict the fth subset using SAS PROC PLM procedure was calculated. This procedure was repeated 5 times such that each subset was used for testing exactly once and an average RMSE of 5-fold cross-validation was calculated by using the formula: SQRT((RMSE12 + To test for replication of the trend of the overall outcome mean over time obtained from the overall GZ set, we applied the same methods as described above to the independent pseudo-panel dataset from Chengdu Blood Center, i.e., overall CD set. To detect the potential effect of family/replacement plateletpheresis donations on the model-based trend of the average apheresis platelet units per donor over time identi ed from the overall GZ and CD sets, we removed the family/replacement plateletpheresis donations in each of the rst 11 cross-sections (all donations after the ban were voluntary) from the overall individual-level data. Then we re-grouped the remaining individuals (81801 and 48768 voluntary plateletpheresis donors remaining from the discovery and replicate sets, respectively)to generate two nested pseudo-panel subsets, named as "voluntary GZ subset"(n = 300 cells) and "voluntary CD subset" (n = 284 cells), respectively. Finally, we used the same methods as described above to t the piecewise linear mixed models for both subsets.
All data management and statistical analyses described above were conducted with R (R Development Core Team) and SAS v9.4 (SAS Institute, Cary, North Carolina).

Results
Demographics of plateletpheresis donors across 14 successive cross-sections with a 6-month interval each in both overall discovery and replicate sets Table 1 and Additional le 2 showed a similar demographics of plateletpheresis donors across 14 successive cross-sections between overall discovery and replicate sets. Our overall discovery set consisted of 69.3-80.3% of plateletpheresis donors who were male across all 14 cross-sections, 5.0-6.5% and 8.8-9.7% ≥46 years old across the rst 11 cross-sections (i.e., before the ban) and the 12th -14th cross-sections (i.e., after the ban), respectively, as well as 45.  Table 1). The overall GZ set also contained about 15.4-41.9% of FRDs across the rst 11 cross-sections (for the 12th -14th cross-sections, no donations were FRD) ( Table 1). In the overall replicate set, about 63.3-75.5% were male across all 14 cross-sections; 10.0-13.9% and 16.3-17.9% were ≥ 46 years old across the rst 11 cross-sections and the 12th -14thcross-sections, respectively;30.1-40.6% and 63.9-72.4% had a blood donation history across the cross-sections before and after the ban, respectively; and 3.9-52.6% were family/replacement donors across the rst 11 cross-sections (Additional le 2).  a"None"=no blood donation history; "WB"=whole blood donation history only; "PLT"=plateletpheresis donation history only; "Both"=both whole blood and plate donations history.
In either discovery or replicate dataset, both average total number of donations and average total apheresis platelet units were signi cantly increase dafter the ban compared to those before the ban; in contrast, the change of average total number of donors from before to after the ban was the opposite(t = 2.42-5.56, p = 0.0325-0.0001) ( Table 2, Additional le 3 and Additional le 4).These results indicated that there is a phenomenon of" fewer donors, more blood" after the FRD ban.   Bold font indicates statistical signi cance for the indicated group vs. other group(s) within the same variable at the α = 0.05 level. Z-test was used to compare two mean differences: Z=(mean difference1-mean difference2)/sqrt(se12 + se22), two-tailed.
a"None"=no blood donation history; "WB"=whole blood donation history only; "PLT"=plateletpheresis donation history only; "Both"=both whole blood and plateletpheresis donations history. Bold font indicates statistical signi cance for the indicated group vs. other group(s) within the same variable at the α = 0.05 level. Z-test was used to compare two mean differences: Z=(mean difference1-mean difference2)/sqrt(se12 + se22), two-tailed.
a"None"=no blood donation history; "WB"=whole blood donation history only; "PLT"=plateletpheresis donation history only; "Both"=both whole blood and plateletpheresis donations history. Model-based trend of average apheresis platelet units per donor over time in overall GZ set Table 4 summarized the characteristics of two independent pseudo-panel datasets (overall GZ set: n = 330 cells; overall CD set: n = 316 cells) with the number of plateletpheresis donors per cell ≥ 30. Figure 1A and Fig. 1B showed24 individual trajectories of the average apheresis platelet units per donor representing 24 cohort groupswith each who shared common gender, birth year, and blood donation history over 14 cross-sections in the overall GZ and CD pseudo-panel sets, respectively. Figure 1C and Fig. 1D visualized an overall mean pro le of the outcome over time in the GZ and CD sets, respectively, indicating that there was a breakpoint at the 11th cross-section (just right before the ban) with a roughly horizontal line to the left of the breakpoint and signi cantly positive linear trend to the right. a"None"=no blood donation history; "WB"=whole blood donation history only; "PLT"=plateletpheresis donation history only; "Both"=both whole blood and plateletpheresis donations history.
bValues in bracket are the minimum and maximum number of individual platelet donors in the cellsacross all involved cross-sections.
Thus, to the overall GZ set, we tted a two-piecewise linear mixed-effects model in which we speci ed intercept as the random term with an unstructured covariance-structure. The covariate birth year was excluded due to its non-signi cance. As shown in Table 5 and Fig. 2A, the nal model presented a twopiecewise linear trend of average apheresis platelet units per donor over time (i.e., cross-section) with a horizontal line to the left of the breakpoint (β timeBefore11 = 0.0111, p = 0.0976) and a signi cantly positive linear trend to the right (β timeAfter11 = 0.0404, p < 0.0001). This result suggests that the average apheresis platelet units per donor were maintained at lower level and did not change with time before the ban, but started increasing linearly with time after the ban.  c"None"=no blood donation history; "WB"=whole blood donation history only; "PLT"=plateletpheresis donation history only; "Both"=both whole blood and plateletpheresis donations history.
Bold values denote statistical signi cance.
In the model, independent covariates measured whether baseline outcome differed by group. Table 5  Due to the non-signi cance of the slope for the timeBefore11 term, we only considered the interactions between the timeAfter11 and covariates in the model. After the model selection, a two-way interaction term gender*blood donation history and a three-way interaction term timeAfter11*gender*blood donation history were excluded due to their non-signi cance. The regression coe cient of the interaction term measured whether the outcome change differed by covariate-speci c groups. As shown in Table 5, on average after the ban, the outcome change was signi cant in males than that in females (β time×male vs time×female = 0.0550, p < 0.0001) and in donors with blood donation history than that in their peers without donation history (β time×WB vs time×none = 0.0331, p = 0.0018; β time×PLT vs time×none = 0.0698, p < 0.0001; and β time×Both vs time×none = 0.0444, p < 0.0001).These results suggest that the contributions of gender and blood donation history to the model had heterogeneity for both the baseline outcome value and the outcome change after the ban.
The Pearson residual plots, either marginally for the consideration of xed effects only or conditionally for the consideration of both xed and random effects, for the nal model indicated that the model's assumptions -normality and heteroscedasticity were not signi cantly violated (Additional le 10). Five-fold cross-validation demonstrated that the over-tting percentage of the nal model for overall GZ set accounted for only 0.28% ( Additional le 6), suggesting that there was no signi cant over-tting issue for the nal model. Replication of the model-based trend of average apheresis platelet units per donor over time in overall CD set Next, we used an independent overall CD set to test the replication of the nal model obtained from the overall GZ set. As shown in Table 5 and Fig. 2B, all parameters' estimates in the nal model for the overall CD set were similar to those for the overall GZ set, except for that the outcome change for the donors with whole blood donation history only was not signi cant and that the outcome values across all time-points appeared to be systematically reduced in the overall CD set, compared to those in the overall GZ set. The latter result was consistent with that from the individual-level data, i.e., the overall average apheresis platelet units per donor in Guangzhou (overall mean = 3.3U, SD = 1.0) was higher than that in Chengdu (overall mean = 2.6U, SD = 1.0) with a marginally non-signi cance (mean difference = 0.7U, 95% CI:-0.04-1.5, p = 0.0644, Additional le 5). The assumptions of normality and heteroscedasticity were not signi cantly violated (Additional le 10), and the over-tting percentage of the nal model for overall CD set was 2.52% (Additional le 6). These results suggest that the two-piecewise linear trend of the average apheresis platelet units per donor over time with the cross-section in which the ban went into effect as a time breakpoint that we observed using overall GZ set can be replicated using an independent dataset -overall CD set with some minor variations.

Discussion
In the present study, we observed that total number of donations and total apheresis platelet units after the ban were signi cantly higher than those before the ban whereas total number of plateletpheresis donors showed the opposite in both discovery and replicate datasets. These ndings quantitatively con rmed an emerging phenomenon of "fewer donors, more blood" that we preliminarily observed in the eld since the implement of the FRD ban in China, which breaks the belief of "more donors, more blood" in the past plateletpheresis donation practice.
Furthermore, using a pseudo-panel data approach, we observed that the average apheresis platelet units per donor presented a horizontal line over time before the ban, and then followed a signi cantly positive linear trend over time after the ban from both discovery and replicate datasets. To our knowledge, this study is the rst to report such a two-piecewise linear trend of the average apheresis platelet units per donor over time before and after a nationwide FRD ban.
Our modeling results also revealed that male donors and donors with plateletpheresis donation history had an increased baseline outcome and a signi cant outcome change over time after the ban. These ndings are consistent with those from our individual-level data, i.e., both male donors and donors with plateletpheresis donation history had signi cantly larger increases of both average apheresis platelet units per donor and average total number of plateletpheresis donations per donor compared to their peer groups (i.e., female donors or donors with other donation history), which may be de ned as improved plateletpheresis donation behavior. Therefore, a possible mechanism underlying the two-piecewise linear trend of the outcome over time before and after the ban could be the FRD ban-related improved plateletpheresis donation behavior. Evidence showed that the improved donation behavior can be related to the increased altruism [21,22] and social responsibility of blood donors [21] whereas repeat donors had a higher return donation rate with altruistic reasons [22] .On the other hand, studies also demonstrated that male donors more frequently donated blood [22,23] and women were less likely to donate blood [24,25] probably due to the pregnancy-and/or lactation-based absence [24]. However, whether the ban can increase altruism and social responsibility of male donors and/or repeat donors is unknown. Another possible mechanism could be directly related to the implement of the FRD ban. Our ndings indicated that compared to voluntary plateletpheresis donors, family/replacement donors donated signi cantly smaller volume of apheresis platelet, and therefore, removing family/replacement donors from the data led to the trend of outcome over time before the ban being changed from a horizontal line to a weak and positive linear line. Taken together, the two-piecewise linear trend of outcome over time may be an integrative result from two ban-related factors: the implement of the FRD ban and the improved donation behavior of male donors and/or donors with platelet donation history after the ban.
Systematic reduction of the outcome for the model as a whole in the samples from Chengdu Blood Center compared to that from Guangzhou Blood Center is consistent with our previous study, in which we reported that the average blood donation volume per resident was higher in Guangdong, whose capital city is Guangzhou, than that in Sichuan ,whose capital city is Chengdu (3.06U for Guangdong vs. 2.56 U for Sichuan) [26] .
Our study has several strengths. The use of an independent external dataset to replicate the ndings from the discovery dataset signi cantly increased the external validity of our ndings. The internal validity of our results was improved by the application of ve-fold cross-validation. As mentioned above, our datasets presented a pseudo-panel structure, thus, the use of a pseudo-panel data approach maximized the reliability and validity of our analysis. Our study also has some limitations. Only three covariates were available for our modeling in both discovery and replicate datasets, thus, we cannot completely rule out the possible confounding effects of the unmeasured covariates. However, by calculating the linear mixed-effects model's R-square values(1-SSE/(SSE + SSR)) [27], about 80-89% of the variance for the outcome variable can be explained by the independent variables that are included in our nal models in overall GZ and CD datasets, respectively ( Table 5). The nationwide FRD ban in China was effective on April 1, 2018, thus, the number of the cross-sections available after the ban was relatively few and further continuously monitoring the trend of the outcome over time after the ban is needed. Our study was a cross-sectional design, thus, couldn't gure out the causal relationship between the covariates and the outcome. Pseudo-panel data approach is an aggregation method, therefore, we lost some information and statistical power.

Conclusion
In conclusion, the signi cant increase of the average apheresis platelet units per donor over time after the FRD ban may be related to the implement of the FRD ban and the improved donation behavior of male donors and/or donors with platelet donation history after the ban. Our ndings suggest that to further increase the plateletpheresis donations in China, a continuous implement of the FRD ban is encouraged and more rigorous blood donation motivations are needed for those donors who are female and who have no donation history. They may also potentially motivate the policymakers in other countries where the FRD for plateletpheresis donation is still legitimate to phase out their FRD strategy and ultimately achieve 100% voluntary plateletpheresis donation. Availability of data and materials The datasets used during this study will be made available by the investigators on reasonable request.

Competing interests
The authors declare that they have no competing interests. Authors' Contribution JC,YF, and GZ conceived and designed the study. S L,HC,YL, and JK were responsible for the data collection used in this study from Guangzhou Blood Center or Chengdu Blood Center. LZ and YL exported the used data from the information system in Guangzhou Blood Center and Chengdu Blood Center, respectively.
YF and XF supervised and coordinated all aspects of the study in Guangzhou Blood Center and Chengdu Blood Center, respectively. GZ supervised all aspects of data analysis. JC and GZ implemented the whole process of data analysis and wrote the rst draft. All authors contributed to data interpretation, critical revision of the manuscript, and approved the nal version of the manuscript. JC is guarantor. The corresponding authors attest that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Figure 1
The trend of average apheresis platelets units per donor over time for overall pseudo-panel datasets. Panels A and B visualize individual cohort group trajectories (outcome vs. time) with each line representing a trajectory of a cohort group who had common gender, birth year, and blood donation history for overall GZ and CD pseudo-panel datasets, respectively; panels C and D present the trend of overall mean outcome over time for overall GZ and CD pseudopanel datasets, respectively. Figure 2