Validation of a method of blood pressure measurement for a study of hypertension in a black African population

Hypertension is a major public health issue in Black Africa. It is also an important factor of cardiovascular risk. To determine the prevalence of hypertension in a large population in Burkina Faso, it is more practicable to use an automatic device for the measurement of BP. Before the start of the study, we tested the reliability of an automatic technique for the measurement of BP in sitting position according to the reference technique. The manometer of reference was a manual, aneroid, and calibrated manometer. The automatic manometer was oscillometric, validated according to the protocol of the British Hypertension Society. The study was held on voluntary healthy persons, or patients hospitalized in the cardiology department of the national university hospital of Ouagadougou (high blood pressure, valvulopathy, cardiac insufficiency). BP measurement was made in a sequential way, with the manual manometer, and with the automatic manometer, by a single observer. Measurement by the manual manometer was based on the auscultation of Korotkoff's murmurs. Systolic blood pressure (SBP) corresponded to phase I, and diastolic blood pressure (DBP) to phase V. Measurement by the automatic manometer was made by reading the BP shown on the device screen. The 10.0 version of the SPSS software was used for data analysis. Statistical tests were concluded with a risk of 0.05. Confidence intervals included 95% of the subjects. The percentage comparison of hypertensive subjects observed in the population by both methods was made with a paired khi2 test. We used Pearson's correlation to quantify the relation between the measures taken using the two methods. In order to quantify the degree of agreement of the two methods, we used the intraclass correlation coefficient (ICC) for quantitative BP measurements, and Kappa's coefficient for qualitative measurements (determination of normotensive or hypertensive subjects). The study was held on 50 black African subjects, with the average age of 38.54 4.83 (18 years-77 years); 55% were male subjects. With the reference method in the sleeping position, the mean SBP and DBP values were respectively 122.60 8.52 and 70.36 5.22 mmHg. The minimal and maximal SBP observed were 80 and 240 mmHg respectively, and the DBP, 30 and 130 mmHg respectively. With the automatic method in the sitting position, the mean values SBP and DBP was 119.88+7.50 and 74.80 4.36 mmHg. SBP minimal was of 75, the maximal of 210 mmHg, and the minimal and maximal DBP was respectively 51 and 121 mmHg. Pearson's correlation coefficients for SBP and DBP between the two methods were statistically different from zero; 0.92 (p<0.001), and of 0.82 (p<0.001) respectively. The ICC was 0.91 for SBP and 0.78 for DBP. Kappa's coefficient was calculated to estimate agreement for the determination of normotensive or hypertensive subjects; among the 50 subjects, 36 were classified normotensive by the two methods, and eight, hypertensive. The differences of classification concerned five hypertensive subjects and one normotensive subject, according to the referenced method. Kappa's coefficient was 0.65. The distribution of the subjects in hypertensive and normotensive did not differ significantly in the two methods (p=0.22). The agreement between the two methods was found very good for the measurement of SBP and DBP; it was good for the determination of an hypertensive or normotensive subject. It is important to test the reliability of a technique of BP measurement before the evaluation of hypertension prevalence in a large population. The reliability of the technique ensures a good estimation of the disease prevalence. It is also important to use statistically adapted tests, to avoid any wrong conclusion as to the reliability of the technique.