Inconsistency in prevalence of hypertension based on self-reports and use of standard tests: Implications for large scale surveys