Quantitative Methods for Health Research E-Book

Nigel Bruce

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 1163

Veröffentlichungsjahr: 2017

Beschreibung

A practical introduction to epidemiology, biostatistics, and research methodology for the whole health care community

This comprehensive text, which has been extensively revised with new material and additional topics, utilizes a practical slant to introduce health professionals and students to epidemiology, biostatistics, and research methodology. It draws examples from a wide range of topics, covering all of the main contemporary health research methods, including survival analysis, Cox regression, and systematic reviews and meta-analysis—the explanation of which go beyond introductory concepts. This second edition of Quantitative Methods for Health Research: A Practical Interactive Guide to Epidemiology and Statistics also helps develop critical skills that will prepare students to move on to more advanced and specialized methods.

A clear distinction is made between knowledge and concepts that all students should ensure they understand, and those that can be pursued further by those who wish to do so. Self-assessment exercises throughout the text help students explore and reflect on their understanding. A program of practical exercises in SPSS (using a prepared data set) helps to consolidate the theory and develop skills and confidence in data handling, analysis, and interpretation. Highlights of the book include:

• Combining epidemiology and bio-statistics to demonstrate the relevance and strength of statistical methods
• Emphasis on the interpretation of statistics using examples from a variety of public health and health care situations to stress relevance and application
• Use of concepts related to examples of published research to show the application of methods and balance between ideals and the realities of research in practice
• Integration of practical data analysis exercises to develop skills and confidence
• Supplementation by a student companion website which provides guidance on data handling in SPSS and study data sets as referred to in the text

Quantitative Methods for Health Research, Second Edition is a practical learning resource for students, practitioners and researchers in public health, health care and related disciplines, providing both a course book and a useful introductory reference. 

Leseprobe

Quantitative Methods for Health Research

A Practical Interactive Guide to Epidemiology and Statistics

Second Edition

Nigel Bruce

The University of Liverpool UK

Daniel Pope

The University of Liverpool UK

Debbi Stanistreet

The University of Liverpool UK

This edition first published 2018 © 2018 John Wiley & Sons Ltd

Edition HistoryJohn Wiley & Sons (1e, 2009)

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Nigel Bruce, Daniel Pope and Debbi Stanistreet to be identified as the authors of this work has been asserted in accordance with law.

Registered Office(s)John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Office9600 Garsington Road, Oxford, OX4 2DQ, UK

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of WarrantyThe contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging-in-Publication DataNames: Bruce, Nigel, 1955– author. | Pope, Daniel, 1969– author. | Stanistreet, Debbi, 1963– Title: Quantitative methods for health research : a practical interactive guide to epidemiology and statistics / by Nigel Bruce, Daniel Pope, Debbi Stanistreet. Description: Second edition. | Hoboken, NJ : Wiley, 2018. | Includes index. | Identifiers: LCCN 2017030435 (print) | LCCN 2017031313 (ebook) | ISBN 9781118665268 (pdf) | ISBN 9781118665404 (epub) | ISBN 9781118665411 (pbk.) Subjects: | MESH: Epidemiologic Methods | Biometry–methods | Biomedical Research author.–methods Classification: LCC RA652.4 (ebook) | LCC RA652.4 (print) | NLM WA 950 | DDC 614.4072--dc23 LC record available at https://lccn.loc.gov/2017030435

Cover Design: Wiley Cover Images: (Crowd) © samxmeg/Gettyimages; (Map) © Arthimedes/Shutterstock

CONTENTS

Preface

About the Companion Website

1 Philosophy of Science and Introduction to Epidemiology

Introduction and Learning Objectives

1.1 Approaches to Scientific Research

1.2 Formulating a Research Question

1.3 Rates: Incidence and Prevalence

1.4 Concepts of Prevention

1.5 Answers to Self-Assessment Exercises Section 1.1

2 Routine Data Sources and Descriptive Epidemiology

Introduction and Learning Objectives

2.1 Routine Collection of Health Information

2.2 Descriptive Epidemiology

2.3 Information on the Environment

2.4 Displaying, Describing, and Presenting Data

2.5 Routinely Available Health Data

2.6 Descriptive Epidemiology in Action

2.7 Overview of Epidemiological Study Designs

2.8 Answers to Self-Assessment Exercises

Note

3 Standardisation

Introduction and Learning Objectives

3.1 Health Inequalities in Merseyside

3.2 Indirect Standardisation: Calculation of the Standardised Mortality Ratio (SMR)

3.3 Direct Standardisation

3.4 Standardisation for Factors Other Than Age

3.5 Answers to Self-Assessment Exercises

4 Surveys

Introduction and Learning Objectives

Resource Papers

4.1 Purpose and Context

4.2 Sampling Methods

4.3 The Sampling Frame

4.4 Sampling Error, Confidence Intervals, and Sample Size

4.5 Response

4.6 Measurement

4.7 Data Types and Presentation

4.8 Answers to Self-Assessment Exercises

5 Cohort Studies

Introduction and Learning Objectives

Resource Papers

5.1 Why Do a Cohort Study?

5.2 Obtaining the Sample

5.3 Measurement

5.4 Follow-Up

5.5 Basic Presentation and Analysis of Results

5.6 How Large Should a Cohort Study Be?

5.7 Assessing Whether an Association is Causal

5.8 Simple Linear Regression

5.9 Introduction to Multiple Linear Regression

5.10 Answers to Self-Assessment Exercises

6 Case–Control Studies

Introduction and Learning Objectives

Resource Papers

6.1 Why do a Case–Control Study?

6.2 Key Elements of Study Design

6.3 Basic Unmatched and Matched Analysis

6.4 Sample Size for a Case–Control Study

6.5 Confounding and Logistic Regression

6.6 Answers to Self-Assessment Exercises

7 Intervention Studies

Introduction and Learning Objectives

Resource Papers

7.1 Why Do an Intervention Study?

7.2 Key Elements of Intervention Study Design

7.3 The Analysis of Intervention Studies

7.4 Testing More-Complex Interventions

7.5 Analysis of Intervention Studies Using a Cluster Design

7.6 How Big Should the Intervention Study Be?

7.7 Intervention Study Registration, Management, and Reporting

7.8 Answers to Self-Assessment Exercises

8 Life Tables, Survival Analysis, and Cox Regression

Introduction and Learning Objectives

Resource Papers

8.1 Survival Analysis

8.2 Cox Regression

8.3 Current Life Tables

8.4 Answers to Self-Assessment Exercises

Note

9 Systematic Reviews and Meta-Analysis

Introduction and Learning Objectives

Resource Papers

9.1 The Why and How of Systematic Reviews

9.2 The Methodology of Meta-Analysis

9.3 Systematic Reviews and Meta-Analyses of Observational Studies

9.4 Reporting and Publishing Systematic Reviews and Meta-Analyses

9.5 The Cochrane Collaboration

9.6 Answers to Self-Assessment Exercises

10 Prevention Strategies and Evaluation of Screening

Introduction and Learning Objectives

Resource Papers

10.1 Concepts of Risk

10.2 Strategies of Prevention

10.3 Evaluation of Screening Programmes

10.4 Cohort and Period Effects

10.5 Answers to Self-Assessment Exercises

Note

11 Probability Distributions, Hypothesis Testing, and Bayesian Methods

Introduction and Learning Objectives

Resource Papers

11.1 Probability Distributions

11.2 Data That Do Not Fit a Probability Distribution

11.3 Hypothesis Testing: Summary of Common Parametric and Non-Parametric Methods

11.4 Choosing an Appropriate Hypothesis Test

11.5 Bayesian Methods

11.6 Answers to Self-Assessment Exercises

Bibliography

Index

EULA

List of Tables

Chapter 1

Table 1.4.1

Chapter 2

Table 2.1.1

Table 2.1.2

Table 2.2.1

Table 2.2.2

Table 2.2.3

Table 2.2.4

Table 2.3.1

Table 2.4.1

Table 2.4.2

Table 2.4.3

Table 2.4.4

Table 2.4.5

Table 2.4.6

Table 2.5.1

Table 2.5.2

Table 2.6.1

Table 2.7.1

Chapter 3

Table 3.1.1

Table 3.2.1

Table 3.2.2

Table 3.3.1

Table 3.3.2

Table 3.3.3

Chapter 4

Table 4.2.1

Table 4.2.2

Table 4.2.3

Table 4.2.4

Table 4.3.1

Table 4.4.1

Table 4.4.2

Table 4.4.3

Table 4.4.4

Table 4.5.1

Table 4.6.1

Table 4.6.2

Table 4.6.3

Table 4.6.4

Table 4.6.5

Table 4.6.6

Table 4.6.7

Table 4.7.1

Table 4.7.2

Table 4.7.3

Table 4.7.4

Table 4.7.5

Table 4.7.6

Chapter 5

Table 5.3.1

Table 5.4.1

Table 5.5.1

Table 5.5.2

Table 5.5.3

Table 5.5.4

Table 5.5.5

Table 5.5.6

Table 5.5.7

Table 5.5.8

Table 5.6.1

Table 5.6.2

Table 5.6.3

Table 5.6.4

Table 5.7.1

Table 5.8.1

Table 5.8.2

Table 5.8.3

Table 5.8.4

Table 5.9.1

Table 5.9.2

Table 5.9.3

Table 5.9.4

Table 5.9.5

Table 5.9.6

Table 5.5.4

Table 5.5.5

Table 5.5.6

Chapter 6

Table 6.3.1

Table 6.3.2

Table 6.3.3

Table 6.3.4

Table 6.3.5

Table 6.3.6

Table 6.4.1

Table 6.4.2

Table 6.4.3

Table 6.5.1

Table 6.5.2

Table 6.5.3

Table 6.5.4

Table 6.5.5

Table 6.5.6

Table 6.5.7

Table 6.5.8

Table 6.5.9

Table 6.5.10

Chapter 7

Table 7.3.1

Table 7.3.2

Table 7.3.3

Table 7.3.4

Table 7.4.1

Table 7.4.2

Table 7.4.3

Table 7.4.4

Table 7.4.5

Table 7.5.1

Table 7.6.1

Table 7.7.1

Table 7.8.1(a)

Table 7.8.1(b)

Chapter 8

Table 8.1.1

Table 8.1.2

Table 8.2.1

Table 8.2.2

Table 8.3.1

Table 8.3.2

Chapter 9

Table 9.1.1

Table 9.1.2

Table 9.1.3

Table 9.2.1

Table 9.3.1

Table 9.3.2

Table 9.3.3

Table 9.4.1

Chapter 10

Table 10.3.1

Table 10.3.2

Table 10.3.3

Table 10.3.4

Table 10.3.5

Table 10.3.6

Chapter 11

Table 11.1.1

Table 11.1.2

Table 11.2.1

Table 11.2.2

Table 11.2.3

Table 11.3.1

Table 11.3.2

Table 11.3.3

Table 11.3.4

Table 11.3.5

Table 11.3.6

Table 11.3.7

Table 11.3.8

Table 11.3.9

Table 11.3.10

Table 11.3.11

Table 11.3.12

Table 11.3.13

Table 11.3.14

Table 11.3.15

Table 11.3.16

Table 11.3.17

Table 11.3.18

Table 11.3.19

Table 11.3.20

Table 11.3.21

Table 11.3.22

Table 11.4.1

Table 11.5.1

List of Illustrations

Chapter 1

Figure 1.2.1

Research fantasy time.

Figure 1.2.2

Research is a process that can usefully be thought of as being cyclical in nature and subject to many influences from both inside and outside the scientific community.

Figure 1.3.1

Period and point prevalence.

Chapter 2

Figure 2.1.1

Process of recording deaths in England and Wales.

Figure 2.4.1

Histogram showing frequency distribution of PM

10

.

Figure 2.4.2

Histogram for PM

10

showing a symmetric frequency distribution.

Figure 2.4.3

Histogram showing the distributions of student weights; in fact, there are two overlapping distributions, one for male students and one for female students – see text for explanation.

Figure 2.4.4

(a) A positively (right) skewed distribution; (b) a symmetric distribution.

Figure 2.4.5

An outlier?

Figure 2.4.6

Mean and median.

Figure 2.4.7

A distribution bunched together (a) or well spread out (b).

Figure 2.4.8

Histogram showing frequency distribution of PM

10

.

Figure 2.4.9

Relative frequency distribution of PM

10

data.

Figure 2.4.10

(a) Frequency histograms for England and Wales (upper panel) and Liverpool (b) Relative frequency histograms for England and Wales (upper panel) and Liverpool (lower panel).

Figure 2.4.11

Scatterplot comparing gestational age (35–42 weeks) with birthweight.

Figure 2.4.12

Scatterplot comparing gestational age (0–50 weeks) with birthweight.

Figure 2.4.13

Example of a non-linear relationship: Casualties to road users on Friday.

Figure 2.4.14

Examples of (a) strong and (b) weak linear relationships. (a) A positive linear relationship between age and the percentage of body fat; (b) a negative linear relationship between exam scores for maths and English.

Figure 2.4.15

Examples of positive (a) and negative (b) linear relationships. (a) A positive linear relationship between age and the percentage of body fat; b) a negative linear relationship between exam scores for maths and English.

Source:

World Health Statistics 2015. Last accessed December 2015.

Figure 2.4.16

Examples of scatterplots for different distributions with correlation coefficients.

Figure 2.4.17

Scatterplots showing examples of non-linear associations.

Figure 2.6.1

Total dietary intake and age-adjusted death rate per 100,000 people.

Source

: Bingham 2004. Reproduced with permission of Nature Publishing Group.

Chapter 4

Figure 4.2.1

An example of simple random sampling of 10 subjects from a population of 60 (see text).

Figure 4.2.2

An example of cluster sampling of Year 3 children in a town. From the 16 primary schools (clusters) in the town, four have been selected. Within each of the selected schools, the children from Year 3 can be surveyed (see text).

Figure 4.2.3

An example of quota sampling.

Figure 4.4.1

Histogram of the sampling distribution for Table 4.4.1 (using the intervals shown in Table 4.4.2).

Figure 4.4.2

Histogram of the sampling distribution with a larger sample size (using the intervals shown in Table 4.4.3).

Figure 4.4.3

The sampling distribution of the sample mean from a population with mean weight 70 kg.

Figure 4.4.4

Sampling distributions of the sampling mean for different sample sizes.

Figure 4.4.5

Weekly gross earnings of all women in the UK 2013.

Source

: IDS Employment Law Brief. Private sector distribution of gross weekly earnings for full-time employees. Accessed 2015.

Figure 4.4.6

Sampling distribution of the sample mean from a skewed population.

Figure 4.4.7

The normal distribution; 68.2% of all values lie between μ − σ and μ + σ (see text).

Figure 4.4.8

The sampling distribution of the sample mean.

Figure 4.6.1

Summary of process for questionnaire development.

Figure 4.7.1

Data on marital status from Natsal-3 presented as (top) bar chart of frequency distribution and (bottom) pie chart of relative frequency distribution.

Figure 4.7.2

(a) Number of children. (b) Number of children per family.

Chapter 5

Figure 5.1.1

Overview of the structure of a cohort study with a 5-year follow-up.

Figure 5.4.1

Hypothetical scenario in a cohort study with undetected changes in exercise behaviour during the follow-up period, which vary by socioeconomic group.

Figure 5.7.1

Conditions required for factors (e.g. lack of exercise, poor diet) to confound an observed association between smoking and IHD.

Figure 5.7.2

(a) Hypothesised pathways for raised COHb to confound the relationship between smoking and low birthweight. (b) Alternative pathways, with raised COHb in causal pathway between smoking and low birthweight.

Figure 5.7.3

(a) Model 1 – Relationships among high coffee consumption, heavy alcohol consumption, and liver cirrhosis; (b) Model 2 – Relationships among radon gas, smoking, and lung cancer; (c) Model 3 – Relationships among damp housing, poverty, and respiratory illness.

Figure 5.8.1

Components for the equation of a straight line.

Figure 5.8.2

Birthweight and gestational age of babies.

Figure 5.8.3

Difference between observed

y

-value and a straight line.

Figure 5.8.4

The regression of birthweight on gestational age.

Figure 5.8.5

The three sums of squares used in assessing the goodness of fit of a regression model.

Source

: Field 2001. Reproduced with permission of (last edition).

Chapter 6

Figure 6.1.1

Overview of the general structure of a case–control study, with an example taken from the New Zealand study (Paper A).

Figure 6.1.2

A hypothetical case–control comparison, investigating the association between exposure to a risk factor (sleeping on back during pregnancy, compared to sleeping on the left side) and late stillbirth.

Figure 6.1.3

Time perspectives most commonly seen in case–control and cohort studies.

Note:

These perspectives do not always apply as explained in the text.

Figure 6.2.1

Diagram summarizing how gestational age could confound the relationship observed between maternal sleeping on the back and late stillbirth.

Figure 6.3.1

A hypothetical case–control comparison, investigating the association between exposure to a risk factor (sleeping on back during pregnancy, compared to sleeping on the left side) and late stillbirth.

Figure 6.3.2

The relationship between OR and RR by occurrence of the outcome being investigated (incidence among the non-exposed population).

Figure 6.5.1

Confounding factors in myocardial infarction.

Chapter 7

Figure 7.1.1

Generalised structure of a randomised control trial with two groups, testing an intervention in comparison with a control and (in this case) with a 2-year follow-up. Randomisation, blinding, and the use of a placebo are examined in Section 7.2.

Figure 7.2.1

Study flow chart (paper A): progress of participants in the trial of oral nicotine therapy as an aid for reduction in cigarette smoking.

Source

: Bollinger 2000. Reproduced with permission of BMJ Publishing Group Ltd.

Figure 7.2.2

Comparison of placebo and active treatment in a two-arm trial.

Figure 7.3.1

Schematic overview of a crossover trial.

Figure 7.4.1

Flow chart showing stages in study protocol and numbers of participants (Figure 1 from Paper B).

Source

: Day 2002. Reproduced with permission of BMJ Publishing Group Ltd.

Figure 7.5.1

SDQ (performance) score for children in four schools; see text for further explanation.

Chapter 8

Figure 8.1.1

Survival data are usually positively skewed.

Figure 8.1.2

Illustration of survival data for a cohort study; see text for further explanation.

Figure 8.1.3

Kaplan–Meier survival curves for cancer trial example.

Figure 8.1.4

Distribution of serum cotinine concentrations among current non-smokers; lifelong non-smokers and former smokers are shown separately (Figure 1 from Paper A).

Source

: Whincup 2004. Reproduced with permission of BMJ Publishing Group Ltd.

Figure 8.1.5

Proportion of men with major CHD by years of follow-up in each smoking group. ‘Light passive’ refers to lowest quarter of cotinine concentration among non-smokers (0–0.7 ng.ml), ‘heavy passive’ to upper three quarters of cotinine concentration combined (0.8–14.0 ng/ml), and ‘light active’ to men smoking 1–9 cigarettes a day (Figure 2 from Paper A).

Source

: Whincup 2004. Reproduced with permission of BMJ Publishing Group Ltd.

Figure 8.1.6

Trial profile (Figure 1 from Paper B).

Source

: Bonner 2010. Reproduced with permission of Elsevier.

Figure 8.1.7

Percentage of patients surviving after date of randomization in control and intervention groups (Figure 2 from Paper B).

Source

: Bonner 2010. Reproduced with permission of Elsevier.

Chapter 9

Figure 9.1.1

Recommendations to use thrombolytic therapy lagged well behind the evidence that would have been available from a systematic review and meta-analysis (adapted from Antman 1992). The letter M indicates that at least one meta-analysis was published that year.

Figure 9.1.2

Flow of information through the different phases of a systematic review.

Figure 9.2.1

Funnel plots from two different systematic reviews.

Figure 9.2.2

Random-effect meta-analysis of household treatment water quality interventions (Figure 2 from Paper A).

Source

: Fewtrell 2005. Reproduced with permission of Elsevier.

Figure 9.2.3

Funnel plot of the 24 trials in which deaths occurred.

Source

: Cochrane Injuries Group Albumin Reviewers 1998. Reproduced with permission of BMJ Publishing Group Ltd.

Figure 9.2.4

Meta-analysis of relative risk of death associated with intervention (albumin) compared with control (no albumin) in critically ill patients.

Source

: Cochrane Injuries Group Albumin Reviewers 1998. Reproduced with permission of BMJ Publishing Group Ltd.

Chapter 10

Figure 10.1.1

YLL and YLD for the 22 sub-regions used in the GBD-2010 study; data are for 1990 (a) and 2010 (b).

Source

: Murray 2012. Reproduced with permission of Elsevier.

Figure 10.1.2

Global disability-adjusted life year ranks with 95% uncertainty intervals (UIs) for the top 25 causes in 1990 and 2010, and the percentage change with 95% UIs between 1990 and 2010.

Source

: Murray 2012. Reproduced with permission of Elsevier.

Figure 10.2.1

Age-specific mortality in men according to blood pressure and age, from life insurance data: (a) relative risk, and (b) absolute risk (Figure 2 from Paper A). Adapted from Rose 1981.

Figure 10.2.2

Prevalence distribution of serum cholesterol concentration related to coronary heart disease mortality in men aged 55–64 years. Number above each bar represents estimate of attributable deaths per 1,000 population per 10 years. Based on Framingham Study (Figure 3 in paper A).

Figure 10.2.3

Schematic representation of distribution of cholesterol in the population, the relative risk (RR) across the range of cholesterol, and a cut-off level of cholesterol indicating high risk (based on Figure 3 from paper A).

Figure 10.2.6

Schematic representation of distribution building on Figure 10.2.3 of cholesterol in the population, the relative risk (RR) across the range of cholesterol, a cut-off level of cholesterol indicating high risk, and a shift of the distribution to the left.

Figure 10.2.5

Distribution of BMI in 13,716 Australian women aged 45–50 years in 1996 (Figure 1 from

Brown et al.

).

Figure 10.3.1

Distributions of disability score for (a) subjects without persistent back pain and (b) subjects with persistent back pain.

Figure 10.3.2

Receiver-operator characteristic (ROC) curve describing the discrimination of the disability score for persistent back pain.

Figure 10.3.3

Illustration of lead-time bias that may occur in evaluation studies of screening programmes.

Figure 10.4.1

Age-standardised suicide rates: 1950–1999, England and Wales (3-year moving averages) in males (Figure 1(a) in paper B).

1

Figure 10.4.2

Suicide and undetermined death rates by time period (of death) and by age group in (a) men and (b) women in the UK (Figure 2 in paper B).

Figure 10.4.3

Rates of male suicide and undetermined death in successive 5-year birth cohorts by age group. The

x

-axis (age-groups) shows age at death (

Figure 3

(a) in paper B).

Figure 10.4.4

Rates of male suicide and undetermined death in successive 5-year birth cohorts by age group, excluding overdose and gassing. The

x

-axis (age-groups) shows age at death (Figure 3(b) from paper B).

Chapter 11

Figure 11.1.1

Frequency distributions of PM

10

concentrations (same data as Chapter 2).

Figure 11.1.2

Probability distribution for obtaining heads after two tosses of a coin.

Figure 11.1.3

Probability distribution for PM

10

data showing probability density.

Figure 11.1.4

Events such as these that occur randomly in time and are not dependent on the timing of other events are described by the Poisson distribution.

Figure 11.1.5

Poisson distribution for increasing values of

µ

.

Source

: Kirkwood 2010. Reproduced with permission of John Wiley & Sons.

Figure 11.2.1

Effect of log transformation of skewed survival data.

Figure 11.2.2

Skewed data on vitamin D levels (left) and natural log transformation (right).

Figure 11.2.3

Common transformations for reducing the degree of skew in distributions.

Figure 11.3.1

Alkaline phosphatase levels in (a) healthy subjects and (b) patients with coeliac disease.

Figure 11.3.2

Distribution of differences in the asthma test scores.

Figure 11.3.3

Box and whisker plot (see text) showing systolic blood pressure for four occupational groups.

Figure 11.3.4

(a) Box-and-whisker plot showing alcohol intake (units per week) by four occupational groups and (b) distribution of alcohol intake (units per week) for administrative workers.

Figure 11.5.1

Fagan's nomogram.

Guide

Cover

Table of Contents

Preface

Pages

iii

iv

xv

xvi

xvii

xviii

xix

xx

xxi

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

60

61

62

63

64

65

66

67

68

69

71

72

73

74

75

76

77

78

79

80

81

82

83

84

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

156

157

158

159

160

161

163

164

165

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

402

403

404

405

406

407

408

409

410

411

413

414

415

416

418

419

422

423

424

425

426

427

429

430

431

432

433

434

435

436

438

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

477

478

479

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

519

520

521

522

523

524

525

526

527

529

530

531

533

534

535

536

537

538

539

540

541

542

543

544

545

Preface

Introduction

Welcome to Quantitative Methods for Health Research, a study programme designed to introduce you to the knowledge and skills required to make sense of published health research, and to begin designing and carrying out studies of your own.

The book is based closely on materials developed and tested over more than 15 years with the campus-based and online Master of Public Health (MPH) programmes at the University of Liverpool, UK. A key theme of our approach to teaching and learning is to ensure a reasonable level of theoretical knowledge (as this helps to provide a solid basis to understanding), while placing at least as much emphasis on the application of theory to practice (to demonstrate what actually happens when theoretical ‘ideals’ come up against reality). For these reasons, the learning materials have been designed around a number of published research studies and information sources that address a variety of topics from around the world, including both developed and developing countries. The many aspects of study design and analysis illustrated by these studies provide examples which are used to help you understand the fundamental principles of good research, and to practise these techniques yourself.

The MPH programme on which this book is based consists of two postgraduate taught epidemiology and statistics modules, one Introductory and the other Advanced, each of which requires 150 hours of study (including assessments), and provides 15 postgraduate credits (1 unit). As students and tutors using the book may find it convenient to follow a similar module-based approach, the content of chapters has been organised to make this as simple as possible. The table summarising the content of each chapter on pages xvii to xix indicates which sections (together with page numbers) relate to the introductory programme, and which to the advanced programme.

The use of computer software for data analysis is a fundamental area of knowledge and skills for the application of epidemiological and statistical methods. A complementary study programme in data analysis using IBM SPSS software has been prepared; this relates closely to the structure and content of the book. Full details of this study programme, including the data sets used for data analysis exercises, are available on the companion website for this book www.wiley.com/go/bruce/quantitative-health-research.

The book also has a number of other features designed to enhance learning effectiveness, summarised in the following sections.

Learning Objectives

Specific, detailed learning objectives are provided at the start of each chapter. These set out the nature and level of knowledge, understanding, and skills required to achieve a good standard at the master's level, and can be used as one point of reference for assessing progress.

Resource Papers and Information Sources

All sections of published studies that are required, in order to follow the text and answer the self-assessment exercises, are reproduced as excerpts in the book. However, we strongly recommend that all resource papers be obtained and read fully, as indicated in the text. This will greatly enhance the understanding of how the methods and ideas discussed in the book are applied in practice, and how research papers are prepared. All papers are fully referenced and available through open-access, in journals that are easily available through higher education establishments.

Key Terms

In order to help identify which concepts and terms are most important, those regarded as core knowledge appear in bold italic font. These can be used as another form of self-assessment, as a good grasp of the material covered in this book will only have been achieved if all these key terms are familiar and understood.

Sample Size Calculations

In several chapters, sample size calculations are explained and used as a basis for self-assessment exercises. We use OpenEpi webtools for this purpose, which can be found at http://www.openepi.com.

SPSS Dataset Used for Illustrating Examples of Statistical Analysis

A reference dataset is used in the book to illustrate analytical output from regression analyses (Chapters 5 and 6) using SPSS. It is also used for other data manipulation and analysis exercises for workbooks located on the companion website for the book (www.wiley.com/go/ bruce/quantitative-health-research). This reference dataset relates to how aspects of work in manual occupational settings are associated with the outcome of low back pain, and has the following features:

The aim of the study was to see what features of the occupational environment were associated with

low back pain

.

The dataset relates to information collected on 775 employees selected randomly from manual occupational settings in the North West of England.

The dataset includes information on

demography

(age, sex, height, weight and social class),

physical working environment

(working postures, manual handling activities and repetitive upper limb movements – the duration of these activities was recorded for 60 minutes of one shift),

psychosocial working environment

(psychological demands of work) and

psychological distress

(a score based on responses to a psychological questionnaire – a higher score indicates a higher level of psychological distress).

Self-Assessment Exercises

Each chapter includes self-assessment exercises, which are an integral part of the study programme. These have been designed to assess and consolidate the understanding of theoretical concepts and competency in practical techniques. The exercises have also been designed to be worked through as they are encountered, as many of the answers expand on issues that are introduced in the main text. The answers and discussion for these exercises are provided at the end of each chapter.

Mathematical Aspects of Statistics

It has been our experience that many students interested in health research, while motivated and very capable, nevertheless do find that the mathematical aspects of statistical methods, such as formulae and mathematical notation, are quite daunting. This is an area of study that does require some persistence, as it is valuable to gain at least a basic mathematical understanding of the most commonly used statistical concepts and methods. We recognise, however, that creating the expectation of much more in-depth knowledge for all readers would be very demanding, and arguably unnecessary. For the most part, therefore, this book avoids detailed mathematical explanation and formulae.

Readers with more affinity for and knowledge of mathematics may be interested to know more, and such understanding is very important for more advanced research work and data analysis. In order to meet these objectives, all basic concepts, simple mathematical formulae, etc., the understanding of which can be seen as core knowledge, are included in the main text. More detailed explanations, including some more complex formulae and examples, are included in statistical reference sections [RS], marked with a start and finish as indicated below.

   RS – Reference Section on Statistical Methods

Text, formulae, and examples.

We hope that you will enjoy this study programme, and find that it meets your expectations and needs.

Organisation of Subject Matter by Chapter

The following table summarises the subject content for each chapter, indicating which sections are introductory and which are advanced.

Chapter content and level

Chapter

Level

Pages

Topics covered

Philosophy of science and introduction to epidemiology

Introductory

1–24

Approaches to scientific research

What is epidemiology?

What is statistics?

Formulating a research question

Rates, incidence, and prevalence

Concepts of prevention

Routine data sources and descriptive epidemiology

Introductory

25–100

Routine collection of health information

Descriptive epidemiology

Information on the environment

Displaying, describing, and presenting data

Association and correlation

Summary of routinely available data relevant to health

Descriptive epidemiology in action, ecological studies, and the ecological fallacy

Overview of epidemiological study designs

Standardisation

Introductory

101–122

Rationale for standardisation

Indirect standardisation

Direct standardisation

Surveys

Introductory

123–184

Rationale for survey methods

Sampling methods

The sampling frame

Sampling error, sample size, and confidence intervals

Response rates

Measurement, questionnaire design, and validity

Data types and presentation: categorical and continuous

Cohort studies

Introductory

185–250

Rationale for cohort study methods

Obtaining a sample

Measurement and measurement error

Follow-up for mortality and morbidity

Basic analysis – relative risk, hypothesis testing (the

t

-test and the chi-squared test)

Introduction to the problem of confounding

Advanced

Sample size for cohort studies

Simple linear regression

Multiple linear regression: dealing with confounding factors

Case–control studies

Introductory

251–296

Rationale for case–control study methods

Selecting cases and controls

Matching – to match or not?

The problem of bias

Basic analysis – the odds ratio for unmatched and matched designs

Advanced

Sample size for case control studies

Matching with more than one control

Multiple logistic regression

Intervention studies

Introductory

297–354

Rationale for experimental study methods

The randomised controlled trial (RCT)

Randomisation

Blinding, controls, and ethical considerations

Analysis of trial outcomes: analysis by intention-to-treat and per-protocol

Paired data and cross-over trials

Advanced

Adjustment when confounding factors are not balanced by randomisation

Sample size for experimental studies

Testing more complex interventions; cluster and non-randomised experimental designs

Factorial design

Multilevel analysis

Trial management and reporting

Life tables, survival analysis, and Cox regression

Advanced

355–384

Nature of survival data

Kaplan–Meier survival curves

Cox proportional hazards regression

Introduction to life tables

Systematic reviews and meta-analysis

Advanced

385–428

Purpose of systematic reviews

Method of systematic review

Method of meta-analysis

Special considerations in systematic reviews and meta-analysis of observational studies

The Cochrane Collaboration

Prevention strategies and evaluation of screening

Advanced

429–476

Relative and attributable risk, population attributable risk, and attributable fraction

High-risk and population approaches to prevention

Measures and techniques used in the evaluation of screening programmes, including sensitivity, specificity, predictive value, and receiver operator characteristic (ROC) curves

Methodological issues and bias in studies of screening programme effectiveness

Cohort and period effects

Probability distributions, hypothesis testing, and Bayesian methods

Advanced

477–528

Theoretical probability distributions

Steps in hypothesis testing

Transformation of data

Paired

t

-test

One-way analysis of variance

Non-parametric tests for paired data, two or more independent groups, and for more than two groups

Spearman's rank correlation

Fisher's exact test

Guide to choosing an appropriate test

Multiple significance testing

Introduction to Bayesian methods

Acknowledgements

The preparation of this book has involved the efforts of a number of people whose support we wish to acknowledge: Francine Watkins for preparing the first section of Chapter 1 ‘Approaches to Scientific Research’; Jo Reeve for preparing Section 5 of Chapter 2; James Higgerson for providing valuable advice for the Current Life Tables section in Chapter 8; Paul Blackburn for assistance with graphics and obtaining permission for the reproduction of the resource papers; Chris West for his advice on statistical methods, and data management and analysis; Nancy Cleave and Gill Lancaster for their invaluable contributions to early versions of the statistical components of the materials; our students and programme tutors who have provided much valued, constructive feedback that has helped to guide the development of the materials upon which this book is based; the staff at Wiley for their encouragement and support; and Chris and Ian at Ty Cam for providing a tranquil refuge in a beautiful part of Denbighshire.

About the Companion Website

Quantitative Methods for Health Research – A Practical Interactive Guide to Epidemiology and Statistics is accompanied by a companion website:

www.wiley.com/go/bruce/quantitative-health-research

The website includes:

SPSS workbooks and datasets

1Philosophy of Science and Introduction to Epidemiology

Introduction and Learning Objectives

In this chapter, we will begin by looking at different approaches to scientific research, how these have arisen, and the importance of recognising that there is no single, right way to carry out investigations in the health field. Rather, we will see that different perspectives can be complementary in providing a more complete understanding of any given issue. We will then go on to explore the research task, discuss what is meant by epidemiology and statistics, and look at how these two disciplines are introduced and developed in the book. The next section introduces the concept of rates for measuring the frequency of disease or characteristics we are interested in, and in particular the terms incidence and prevalence. These definitions and uses of rates are fundamental ideas with which you should be familiar before we look in more detail at research methods and study design. In the final section, we will look at key concepts in disease prevention, including the commonly used terms primary, secondary, and tertiary prevention.

The reason for starting with a brief exploration of the nature of scientific methods is to see how historical and social factors have influenced the biomedical and social research traditions that we take for granted today. This will help you understand your own perceptions of, and assumptions about, health research, based on the knowledge and experience you have gained to date. It will also help you understand the scientific approach being taken in this book and how this both complements and differs from that developed in books and courses on qualitative research methods – as and when you may choose to study these. Being able to draw on a range of research traditions and their associated methods is especially important for the discipline of public health, but it is also important for many other aspects of health and health care.

Learning Objectives

By the end of this chapter, you should be able to do the following:

Briefly describe the key differences between the main approaches to research that are used in the health field.

Describe what is meant by epidemiology, and list the main uses to which epidemiological methods and thought can be put.

Describe what is meant by statistics, and list the main uses to which statistical methods and thought can be put.

Define and calculate rates, prevalence, and incidence, and give examples of their use.

Define primary, secondary, and tertiary prevention and give examples of each.

1.1 Approaches to Scientific Research

1.1.1 History and Nature of Scientific Research

Scientific research in health has a long history going back at least to the classical period. There are threads of continuity, as well as new developments in thinking and techniques, that can be traced from the ancient Greeks and through the fall of the Roman Empire, the Dark Ages, and the Renaissance to the present time. At each stage, science has influenced, and has been influenced by, the culture and philosophy of the time. Modern scientific methods reflect these varied historical and social influences. So it is useful to begin this brief exploration of scientific health research by reflecting on our own perceptions of science and how our own views of the world fit with the various ways research can be approached. As you read this chapter you might like to think about the following questions:

What do you understand by the terms

science

and

scientific research

, especially in relation to health?

How has your understanding of research developed?

What type of research philosophy best fits your view of the world and the health issues you are most interested in?

Thinking about the answers to these questions will help you understand what we are trying to achieve in this section and how this can best support the research interests that you have and are likely to develop in the years to come. The history and philosophy of science is of course a whole subject in its own right, and this is of necessity a very brief introduction.

Scientific Reasoning and Epidemiology

Health research involves many different scientific disciplines, many of which you will be familiar with from previous training and experience. Here we are focusing principally on epidemiology, which is concerned with the study of the distribution and determinants of disease within and between populations. In epidemiology, as we shall see, there is an emphasis on empiricism, that is, the study of observable phenomena by scientific methods, detailed observation, and accurate measurement. The scientific approach to epidemiological investigation has been described as

Systematic

– There is an agreed system for performing observations and measurement.

Rigorous

– The agreed system is followed exactly as prescribed.

Reproducible

– All the techniques, apparatus, and materials used in making the observations and measurements are written down in enough detail to allow another scientist to reproduce the same process.

Repeatable

– Scientists often repeat their own observations and measurements several times in order to increase the reliability of the data. If similar results are obtained each time, the researcher can be more confident the phenomena have been accurately recorded.

These are characteristics of most epidemiological study designs and are an important part of the planning and implementation of the research. However, this approach is often taken for granted by many investigators in the health field (including epidemiologists) as the only way to conduct research. Later we will look at some of the criticisms of this approach to scientific research, but first we need to look in more detail at the reasoning behind this perspective.

Positivism

The view of science and knowledge known as positivism is the dominant philosophy underlying contemporary epidemiology. The evolution of positivism has been extensively documented elsewhere (see Guba and Lincoln, 1994; Halfpenny, 1982; and Feigl, 1969), its early development being attributed mainly to August Comte during the early 19th century. However, its roots can be traced back to the 17th century, to a time when scientists stopped relying on religion, conjecture, and faith to explain phenomena, and instead began to use reason and rational thought. This period saw the emergence of the view that it is only by using scientific thinking and practices that we can reveal the truth about the world.

Positivism assumes a stable observable reality that can be measured and observed. So, for positivists, scientific knowledge is proven knowledge, and theories are therefore derived in a systematic, rigorous way from observation and experiment. This approach to studying human life is the same approach that scientists take to study the natural world. Human beings are believed by positivists to exist in causal relationships that can be empirically observed, tested, and measured (Bilton et al., 2002) and to behave in accordance with various laws. Because this reality exists whether we look for it or not, it is the role of scientists to reveal its existence but not to attempt to understand the inner meanings of these laws or express personal opinions about these laws. One of the primary characteristics of a positivist approach is that the researcher takes an objective distance from the phenomena so that the description of the investigation can be detached and undistorted by emotion or personal bias (Davey, 1994). This means that within epidemiology, various study designs and techniques have been developed to increase objectivity; you will learn more about these in later chapters.

More recently, some of the earlier tenets of positivism have been challenged through the work of Karl Popper and other scholars such as Bronowski (Bronowski, 1956; Popper, 1959), and as a result, a post-positivistic approach has emerged since the mid-20th century or so, and this approach now underpins much contemporary empirical research activity (Philips, 1990). Post-positivism still advocates that there is an objective reality, but it suggests that reality can only be measured imperfectly due to the limitations of the scientific approach. It also asserts a realist perspective, stating that there are phenomena that can't be observed but nevertheless do exist, so science is not limited to only those phenomena that can be measured. A more-detailed discussion of post-positivism is outside of the scope of this book. For the purposes of this chapter, when we refer to positivism, this can be assumed to refer to both positivist and post-positivist approaches within epidemiology.

A second aspect of scientific thinking, which also evolved over this period, derives from the work of Thomas Kuhn (1922–1996), who challenged the concept of absolute evidence. In The Structure of Scientific Revolutions, (Kuhn, 1970), Kuhn argued that one scientific paradigm – one ‘conceptual worldview’ – may be dominant at a particular period in history. Over time, however, this paradigm is challenged, and eventually it is replaced by another view (paradigm), which then becomes accepted as the most important and influential. He termed these revolutions in science ‘paradigm shifts’. Although questioned by other writers, this perspective suggests that scientific methods we may take for granted as being the only or best way to investigate health and disease are to an extent the product of historical and social factors, and they can be expected to evolve – and maybe change substantively – over time.

Induction and Deduction

There are two main forms of scientific reasoning: induction and deduction. Both have been important in the development of scientific knowledge, and it is useful to appreciate the difference between the two in order to understand the approach taken in epidemiology.

Induction

With inductive reasoning, researchers make repeated observations and use this evidence to generate theories to explain what they have observed. For example, if a researcher made a number of observations in different settings of women cooking dinner for their partners, they might then inductively derive a general theory:

All women cook dinner for their partners.

Deduction

Deduction works in the opposite way to induction, starting with a theory (known as an hypothesis) and then testing it by observation. Thus, a very important part of deductive reasoning is the formulation of the hypothesis – that is, the provisional assumption researchers make about the population or phenomena they wish to study before starting with observations. A good hypothesis must enable the researcher to test it through a series of empirical observations. So, in deductive reasoning, the hypothesis would be

All women will cook dinner for their partners.

Observations would then be made in order to test the validity of this statement. This would allow researchers to check the consistency of the hypothesis against their observations, and if necessary, the hypothesis can be discarded or refined to accommodate the observed data. So, if they found even one woman not cooking for her partner, the hypothesis would have to be re-examined and modified. This characterises the approach taken in epidemiology and by positivists generally.

Karl Popper (1902–1994) argued that hypotheses can never be proved true for all time, and scientists should aim to refute their own hypotheses even if this goes against what they believe (Popper, 1959). He called this the hypothetico-deductive method, and in practice this means that an hypothesis should be capable of being falsified and then modified. Thus, to be able to claim the hypothesis is true would mean that all routes of investigation have been carried out. In practice, this is impossible, so research following this method does not set out with the intention of proving that an hypothesis is true. In due course we will see how important this approach is for epidemiology and in the statistical methods used for testing hypotheses.

Alternative Approaches to Research

It is important to be aware that positivism is only one of many different approaches to scientific research. Many social scientists, for example, believe that these approaches are not relevant for the study of human behaviour. From this perspective, they believe that human beings do not act in accordance with observable rules or laws. This makes humans different from phenomena in the natural world, and so they need to be studied in a different way. Positivism (and post-positivism) have also been criticised because they cannot explain how people interpret or make sense of the world. As Green and Thorogood (2004, p. 12) argue,

Unlike atoms (or plants or planets), human beings make sense of their place in the world, have views on researchers who are studying them, and behave in ways that are not determined in law-like ways. They are complex, unpredictable, and reflect on their behaviour. Therefore, the methods and aims of the natural sciences are unlikely to be useful for studying people and social behaviour: instead of explaining people and society, research should aim to understand human behaviour.

Many social scientists therefore hold different beliefs about how we should carry out research into human behaviour. Consequently, they are more likely to take an inductive approach to research because they argue that they do not want to make assumptions about the social world until they have observed it in and for itself. They therefore do not want to formulate hypotheses because they believe these are inappropriate for making sense of human action. Rather, they believe that human action cannot be explained but must be understood.

Whereas positivists would be concerned mainly with observing patterns of human behaviour, other researchers principally wish to understand that behaviour. This latter group requires a different starting point that will encompass their view of the world, or different theoretical positions to make sense of the world. It turns out that there are many different positions that can be adopted, and while we cannot go into them all here, we briefly consider one of the most important of these, known as an interpretative approach.

An interpretative approach assumes an interest in the meanings underpinning human action, and the role of the researcher is therefore to unearth that meaning. The researcher would not look to measure the reality of the world but would seek to understand how people interpret the world around them (Green and Thorogood, 2004).

Let's look at an example of positivist and interpretivist approaches in respect of a common health problem with multiple physiological, social, and behavioural aspects, namely, asthma. A positivist approach to researching this condition may be to obtain a series of objective measurements of symptoms and lung function using a standard procedure on a particular sample of people over a specified period of time. An interpretative approach might involve talking in-depth to a small number of participants with asthma to try to understand how they view the impact of their symptoms on their lives. Obviously, in order to do this, these two types of approaches require the use of different research methods. Those planning interpretative research would use qualitative methods (e.g. interviews, focus groups, and ethnographic methods), whereas positivists (e.g. epidemiologists) would choose quantitative methods (e.g. surveys and cohort studies involving lung-function measurements and highly structured questionnaires). These two different approaches would draw on different research paradigms and would therefore produce different types of findings.

Those drawing on an interpretative perspective would also differ from positivists in respect of the view that researchers can have an objective, unimpaired, and unprejudiced stance in the research that allows them to make value-free statements. Interpretative research accepts that researchers are human beings and therefore cannot stand objectively apart from the research. In a sense, they are part of the research process, as their presence can influence the nature and outcome of the investigation.

With the asthma example, we can see how complementary the findings of these two different approaches to research could be. The positivist approach can help determine whether a new medication provides any benefit in terms of control of symptoms or lung function, for example. On the other hand, the interpretative approach can help us understand why the activities of some people may be more affected by their condition than others, for example.

Researchers working with a post-positivist framework have to some extent narrowed the divide between quantitative and qualitative approaches to research, since post-positivism does not reject approaches that focus on the meanings people give to their actions as seen in interpretative approaches to research. Mixed-methods approaches (using both qualitative and quantitative methods) to research can therefore help build up a more-complete picture of effective health care and support that allows a better understanding of many aspects of a person's life and health experience. Further discussion of mixed methods is outside of the scope of this book; for a useful guide, see Cresswell and Piano Clark (2011).

Exercise for Reflection

Make brief notes on the type of scientific knowledge and research with which you are most familiar.

Is this predominantly positivistic (hypothetico-deductive) or interpretative in nature, or is it more of a mixture?

There are no answers provided for this exercise, as it is intended for personal reflection.

1.1.2 What is Epidemiology?

The term epidemiology is derived from the following three Greek words:

Epi

– among

Demos

– the people

Logos

– study of

We can translate this in more modern terms into the study of the distribution and determinants of disease frequency in human populations. The following exercise will help you to think about the uses to which the discipline of epidemiology is put.

   Self-Assessment Exercise 1.1.1

Make a list of some of the applications of epidemiological methods and thought that you can think of. In answering this, avoid listing types of epidemiological study that you may already know. Try instead to think in general terms about the practical applications of these methods.

Answers in Section 1.5

This exercise shows the very wide application of epidemiological methods and thought. It is useful to distinguish between two broad functions of epidemiology, one very practical, the other more philosophical:

The range of epidemiological research methods provides a toolbox for obtaining the best scientific information in a given situation (assuming, that is, you have established that a positivist approach is most appropriate for the topic under study!).

Epidemiology helps us use knowledge about the population determinants of health and disease to inform the full range of investigative work, from the choice of research methods, through analysis and interpretation, to the application of findings to policy. With experience, this becomes a way of thinking about health issues over and above the mere application of good methodology.

You will find that your understanding of this second point grows as you learn about epidemiological methods and their application. This is because epidemiology provides the means of describing the characteristics of populations, comparing them, and analysing and interpreting the differences, as well as the many social, economic, environmental, behavioural, ecological, and genetic factors that determine those differences.

1.1.3 What are Statistics?