Abstract
Sources of error in recording the blood pressures of patients with hypertension in general practice were examined for 1072 patients over two and a half years. A highly significant preference for terminal digit 0 was shown, but terminal digit preference operates in a different manner around important thresholds. Doctors did not adhere consistently to the protocol that they accepted. These sources of bias affect the calculations of mean blood pressure and have considerable implications for decisions about treatment for many patients.
Full text
PDFPage 1661-1664
