Statistical analysis of heart rate, blood pressure and respiration in healthy man and post-operative patient

Abstract

This thesis is concerned with the problem of obtaining a full description of the physiological variables of heart rate and blood pressure. The results should then be useful in the design of machines that measure or monitor these variables, and for research workers 'who wish to analyse data derived from these variables.

The original impetus for this thesis came from a desire by Mr. (now Professor) D. E. IMI. Taylor of the Department of Physiology, Edinburgh to improve the design of a patient monitoring device (PAWS). A large amount of data had been collected from post-operative patients in a coronary care unit. The electrocardiogram and arterial blood pressure had been recorded continuously for periods of. upto an hour, and from these were derived the beat-by-beat heart rate and the mean arterial blood pressure. A description'cf the variability and stability of these two parameters would contribute to the design of a patient monitoring device and should improve the ability of the device to discriminate between normal and abnormal situations.

It is well known that respiration can, on occasions, be a major source of variability of heart rate and blood pressure,, and a separate study was carried out to investigate the effect of respiration or. heart rate. The effects of respiration on blood pressure were not studied since this would have required catheterization of healthy subjects.

Further data were kindly supplied by Professor B. McA. Sayers of Imperial College, London ( Ref. E. E. C. Contract No. 297/76/12 ECI UK. ) These consisted of the arterial systolic and end-diastolic pressures for. each heart beat from two healthy subjects. A two hour stretch of record was examined and during the monitoring the subjects conducted normal activities, hence the term 'ambulatory' monitoring.

The questions that we are going to ask are concerned with the structure and stability of the data. Clearly the sampling distributions of the variables need to be examined, and the nature and properties of the outliers. We also wish to know how the sequential nature of. the data affects the amount of information contained in it and to, examine the sampling distribution of the mean and variance so that the effect of averaging can be estimated. Stability of the data is another aspect to be examined for how long does the blood pressure, for example, remain within defined limits and on average 'hot-i long does it remain*stationary? Further useful information can be obtained by studying the frequency distribution of the data. From this we can estimate the contribution to the variability of various frequency bands, for example the variability due the the respiratory effect and from activities such as the vasomotor response,

The structure of this thesis is as follows: Chapter I is concerned. with data collection and preprocessing, and with the detection and analysis of outliers. In Chapter 2 we examine the stability and stationarity of the patient and ambulatory monitoring data and, in more detail, the structure of the patient data, Chapter 3 is concerned with the theoretical problems associated with the spectral analysis of unequally spaced observations of which heart beats are an obvious example, in preparation for Chapter 4, where the analysis of t1ho data in the frequency domain is considered. Chapter 5 deal with the effect of respiration on heart rate and blood pressure and Chapter 6 discusses modelling aspects of the control systems involved. A summary and are conclusion / given in Chapter 7. Appendix A deals with the mathematical and practical aspects of spectral analysis and digital filtering and Appendix B gives a physiological, background to the central and reflex control of heart rate gnd blood pressure. A description of the computer programs is given in Appendix C. The methods or analysis are described in detail, in the hope that the thesis may prove useful to workers wishing to apply time series methods to records of heart rate, blood pressure and respiration.