We appreciate Dr O'Connors interest in our study, but do not agree with his comments.1
Studies should be sized for their purpose. In this case, 163 participants provided reasonably narrow confidence intervals around the coefficients of variation (CV) estimate, for example; for systolic office blood pressure (BP) 8.6% (95% CI = 7.6 to 9.6%). We feel this interval excludes any meaningfully different clinical interpretation.
Regarding participant selection, few studies select people randomly from the population. Studies of BP almost always select people whose BP is considered clinically relevant – typically, with established hypertension, cardiovascular disease, or risk factors, as in this study. In clinical practice, most patients for whom accurate BP measurement is thought desirable will fall into one of these categories.
The Boso Medicus and Boso Medicus Prestige devices are rebrandings of the A&D UA–767 and UA–787 devices (Welte W, Bosch & Sohn GmbH. Personal communication, 2010), that are listed on the British Hypertension Society website.2 We apologise for this omission.
The CV is the standard way of reporting variability because it allows comparison of variability between samples with different means for example, hypertensive and non-hypertensive patients.
It is not clear how Dr O'Connor prefers the trial data to be explained. We could present the actual BP measurements for individual participants, but BP could be made to appear more or less variable simply by selecting participants with labile or relatively stable BP. The use of a summary statistic for the entire dataset is inevitable if such bias is to be avoided. We felt some interpretation of the CV would be helpful, and this can be applied to real patients as well as hypothetical ones.
Our results are in line with other studies as referenced in our paper.3 We can provide further reassurance from our unpublished analyses of other datasets, that have yielded CV estimates between 7.4% (for a well-known trial with several thousand uncomplicated participants) and 11.6% (for complex patients with imperfect real-life BP measurement). The variability of office BP quoted in this study is likely to be rather conservative, as many clinicians do not have time to measure BP in triplicate after 5-minutes at rest.
REFERENCES
- 1.O'Connor R. Self-monitored blood pressure measurements. Br J Gen Pract. 2011;61(582):63. doi: 10.3399/bjgp11X549063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.British Hypertension Society. Automatic blood pressure measuring devices suitable for use in the clinic and also at home for self-monitoring. Leicester: British Hypertension Society; http://bhsoc.org/bp_monitors/automatic.stm (accessed 11 Jan 2011) [Google Scholar]
- 3.Warren RE, Marshall T, Padfield PL, Chrubasik S. Variability of office, 24-hour ambulatory, and self-monitored blood pressure measurements. Br J Gen Pract. 2010;60(578):675–680. doi: 10.3399/bjgp10X515403. [DOI] [PMC free article] [PubMed] [Google Scholar]