Main text

We thank Chatterjee and Pradhan for their letter regarding our paper in Trials. 2016; 17-517. We agree with their sentiment that local HbA1c measurement cannot be implemented at the expense of clinically unacceptable disparities between centralised and localised measurements despite its greater cost efficiency. We hope that the following provides the additional information that will aid their assessment of our results.

Time-lag

In their letter, Chatterjee and Pradhan draw attention to Fig. 3 of our paper, and notice that within centres the distribution of differences is centred on 0 (suggesting that there are no centre-specific systematic biases present). We would argue that this does not imply that the same relationship would necessarily be true for time-lag. Figure 3 displayed results by site; however, within a site the time lag may vary. We have produced boxplots to show the distribution of differences by time-lag and a scatterplot as requested. This demonstrates an absence of a linear relationship between time-lag and discrepancy. They also indicate that in practice time-lag can be an important factor for high glucose values. As part of the underlying assumptions of the Bland-Altman method we investigated heteroscedasticity (see the Verification of assumptions section within the ‘Results’ section of our paper), i.e. we did not observe any increase in discrepancy with higher glycosylated haemoglobin (HbA1c).

Of the 590 measurements analysed for agreement, for 79 (13.4%) the date of measurement at the central laboratory was not recorded. These 79 are indicated with ‘M’ on Fig. 1 and excluded from Fig. 2 (scatterplot). For the remaining 511, in only 8 (1.5%) measurements was there a time-lag of more than 7 days. The Pearson’s correlation between time-lag and difference in measurements (local minus central HbA1c) was found to be −0.02. This was statistically not different from 0 (p = 0.48). This means that there is no evidence of a straight-line relationship (linear correlation) between time-lag and agreement. (See Figs. 1 and 2)

Fig. 1
figure 1

Boxplots showing the distribution of differences between local and central measurements by time-lag in days: between blood samples being taken and analysis at the central laboratory. M: measurements where the date of laboratory measurement was not recorded

Full size image
Fig. 2
figure 2

Scatterplot showing the distribution of the magnitude of differences between local and central measurements by time-lag in days: between blood samples being taken and analysis at the central laboratory

Full size image

HbA1c measurement methodology

We specify in our paper that in almost all cases, both local and central, HbA1c was measured via immunoassay using portable machines. Local measurements were normally taken at outpatient clinics – but the technical method of measurement employed was not recorded. At the time of analysis, we contacted all sites to establish what the local methodology was. Table 1 gives details of what the sites’ responses were.

Table 1 Local HbA1c measurement methodology, as reported by sites
Full size table

We can give further detail here to say that the same portable machine was used at Alder Hey for outpatient clinics as at the central laboratory (based at Alder Hey). At this centre therefore, the methodology was identical. Differences were still incurred despite using the same machine with a short time-lag and removing the courier and post-transfer issues (see Fig. 3).

Fig. 3
figure 3

Scatterplot of discrepancies in measurements by time-lag at Alder Hey Hospital (location of central laboratory)

Full size image

Whether central laboratory results should be used in preference to local results is an issue that needs to be considered at the design stage of any study. We hope that the information presented will enable greater clarity in decisions made. However, any decision needs to be born against the size of the effect that is to be detected and the potential size of discrepancies. This study demonstrates that despite quality control placed on local machines such discrepancies do occur. It should also be emphasised that this study took place in the UK and the climate and transport conditions elsewhere may determine whether local measurements are preferable.