CD Deakin (Southampton)

Statistician's report by Julie Morris
 

This small observational study of hypotensive patients looks at the presence or absence of radial, femoral and carotid pulses associated with differing systolic blood pressure. The study appears to have been well carried out and the main conclusions are valid. The results do seem at odds with the ATLS course teaching information. However, the presentation of the results needs to be improved.
 

Design and Methods:

1. No justification is given for the small sample size of 20. The number of patients with data for certain pulse/blood pressure combinations is even smaller. In addition, it is not clear whether they were consecutive patients, that is whether a 100% response rate was attained.
2. To use the term, "correlation", is rather misleading because it has a particular statistical meaning and should be applied to continuous or ordinal data alone. It would be more appropriate to talk about the "relationship" or "association" between presence/absence of pulse with a variation in blood pressure.
3. The choice of the 50% centile as the ‘cut-off’ point seems arbitrary. Statements such as, "Using 50% centiles, carotid, femoral and radial pulses are present in patients with a systolic > 70 mmHg…" (Abstract and Discussion, Page 6), are rather meaningless. It would be more appropriate to give the % of patients with pulses present for particular values/ranges of blood pressure together with the associated 95% confidence intervals. The latter would portray the accuracy of these estimates for this small sample of patients.
4. Table 1 could be omitted.