Applying 10 different methods to handle outliers in ambulatory blood pressure monitoring resulted in discarded data varying from 1% to 17%, strongly reducing the standard deviation of readings.
Does the choice of method to handle outliers in ambulatory blood pressure monitoring significantly affect the resulting mean values and standard deviations?
Different methods for handling outliers in ambulatory blood pressure monitoring yield significantly different standard deviations, highlighting the need for a standardized approach.
OBJECTIVES: To examine the methods to handle marginal readings in the analysis of ambulatory blood pressure. DESIGN: Data obtained from automatic ambulatory blood pressure monitoring include several 'outliers', i.e. readings at the frontier of physiologically acceptable ranges. Several methods have been used to handle these readings. We need to know whether using different methods to reject outliers leads to different results in the analysis of the data. If so, then it is important that a common method be used by different authors. METHODS: Ten reported methods to handle outliers were selected and applied to a large set of unpublished blood pressure profiles (Novacor Diasys system). We compared the effects of data rejection by these methods on the mean values and standard deviations (calculated over 24 h, daytime and night-time) of the remaining data. RESULTS: The different methods had quite different effects on the same data set. Depending on the method used, the discarded data varied from 1 to 17% of the total number of readings. Among the rejected data, readings that occurred in the daytime varied from 14 to 56%. Also, 'high-value' outliers varied from 1 to 60% of the rejected data. On average over the total sample, the rejection of outliers had only a small effect on the mean values of blood pressure. In contrast, it may strongly reduce the standard deviation of the readings. CONCLUSION: The study emphasized the need to use a common method to handle outliers in the analysis of ambulatory blood pressure data.
Berardi et al. (Thu,) conducted a review in Ambulatory blood pressure monitoring. 10 reported methods to handle outliers was evaluated on Effects of data rejection on mean values and standard deviations of blood pressure readings. Applying 10 different methods to handle outliers in ambulatory blood pressure monitoring resulted in discarded data varying from 1% to 17%, strongly reducing the standard deviation of readings.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: