Cover Page

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










Wiley Logo

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:

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.

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
  1. 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
  1. 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
  1. Standardisation
Introductory 101–122
  • Rationale for standardisation
  • Indirect standardisation
  • Direct standardisation
  1. 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
  1. 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
  1. 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
  1. 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
  1. 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
  1. 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
  1. 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
  1. 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: