Cover Page







To

Tessa and Emma Fayers

and

Christine Machin

Quality of Life

The assessment, analysis and reporting
of patient-reported outcomes

Third Edition



PETER M. FAYERS

Institute of Applied Health Sciences, University of Aberdeen School of Medicine and Dentistry, Scotland, UK and
Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology (NTNU), Trondheim, Norway


and

DAVID MACHIN

Medical Statistics Group, School of Health and Related Research, University of Sheffield, Sheffield, UK and
Department of Cancer Studies and Molecular Medicine, University of Leicester, Leicester, UK










Wiley Logo

Preface to the third edition

When the first edition of this book was published in 2000, the assessment of quality of life (QoL) as an important outcome in clinical trials and other research studies was, at best, controversial. More traditional endpoints were the norm – measures such as disease status, cure and patient’s survival time dominated in research publications. How times have changed. Nowadays it is generally accepted that the patients’ perspective is paramount, patient representatives are commonly involved in the design of clinical trials, and patient-reported outcomes (PROs) have become recognised as standard outcomes that should be assessed and reported in a substantial proportion of trials, either as secondary outcomes or, in many instances, as a primary outcome from the study. Indeed, in 2000 the term ‘patient-reported outcome’ hardly existed and the focus at that time was on the ill-defined but all embracing concept of ‘quality of life’. Now, we regard QoL as but one PRO, with the latter encompassing anything reported by ‘asking the patient’ – symptoms such as pain or depression, physical or other functioning, mobility, activities of daily living, satisfaction with treatment or other aspects of management, and so on. Drug regulatory bodies have also embraced PROs and QoL as endpoints, while at the same time demanding higher standards of questionnaire development and validation.

In parallel with this, research into instrument development, validation and application continues to grow apace. There is increasing recognition of the importance of qualitative methods to secure a solid foundation when developing new instruments, and a corresponding rigour in applying and reporting qualitative research. In parallel, a major radical shift towards using item response theory both as a tool for developing and validating new instruments and as the basis of computer-adaptive tests (CATs). Many of the major research groups have been developing new CAT instruments for assessing PROs, and this new generation of questionnaires are becoming widely available for use on computer tablets and smart-phones.

Analysis, too, has benefited in various ways for the increased importance being attached to PROs – two examples being (i) methods for handling missing data and in particular reducing the biases that can arise when data are missing, and (ii) greater rigour demanded for the reporting of PROs.

As a consequence of these and many other developments, we have taken the opportunity to update many chapters. The examples, too, have been refreshed and largely brought up-to-date, although some of the classic citations still stand proud and have been retained. A less convenient aspect of the changes is, perhaps, the resultant increase in page-count.

We continue to be grateful to our many colleagues – their continued encouragement and enthusiasm has fuelled the energy to produce this latest edition; Mogens Groenvold in particular contributed to the improvement of Chapter 3.

Peter M. Fayers and David Machin
September 2015

Preface to the second edition

We have been gratified by the reception of the first edition of this book, and this new edition offers the opportunity to respond to the many suggestions we have received for further improving and clarifying certain sections. In most cases the changes have meant expanding the text, to reflect new developments in research.

Chapters have been reorganised, to follow a more logical sequence for teaching. Thus sample size estimation has been moved to Part C, Clinical Trials, because it is needed for trial design. In the first edition it followed the chapters about analysis where we discussed choice of statistical tests, because the sample size computation depends on the test that will be used.

Health-related quality of life is a rapidly evolving field of research, and this is illustrated by shifting names and identity: quality of life (QoL) outcomes are now also commonly called patient- (or person-) reported outcomes (PROs), to reflect more clearly that symptoms and side effects of treatment are included in the assessments; we have adopted that term as part of the subtitle. Drug regulatory bodies have also endorsed this terminology, with the USA Food and Drug Administration (US FDA) bringing out guidance notes concerning the use of PROs in clinical trials for new drug applications; this new edition reflects the FDA (draft) recommendations.

Since the first edition of this book there have been extensive developments in item response theory and, in particular, computer-adaptive testing; these are addressed in a new chapter. Another area of growth has been in systematic reviews and meta-analysis, as evinced by the formation of a Quality of Life Methods Group by the Cochrane Collaboration. QoL presents some particular challenges for meta-analysis, and this led us to include the final chapter.

We are very grateful to the numerous colleagues who reported finding this book useful, some of whom also offered constructive advice for this second edition.

Peter M. Fayers and David Machin
June 2006

Preface to the first edition

Measurement of quality of life has grown to become a standard endpoint in many randomised controlled trials and other clinical studies. In part, this is a consequence of the realisation that many treatments for chronic diseases frequently fail to cure, and that there may be limited benefits gained at the expense of taking toxic or unpleasant therapy. Sometimes therapeutic benefits may be outweighed by quality of life considerations. In studies of palliative therapy, quality of life may become the principal or only endpoint of consideration. In part, it is also recognition that patients should have a say in the choice of their therapy, and that patients place greater emphasis upon non-clinical aspects of treatment than healthcare professionals did in the past. Nowadays, many patients and patient-support groups demand that they should be given full information about the consequences of their disease and its therapy, including impact upon aspects of quality of life, and that they should be allowed to express their opinions. The term quality of life has become a catch-phrase, and patients, investigators, funding bodies and ethical review committees often insist that, where appropriate, quality of life should be assessed as an endpoint for clinical trials.

The assessment, analysis and interpretation of quality of life relies upon a variety of psychometric and statistical methods, many of which may be less familiar than the other techniques used in medical research. Our objective is to explain these techniques in a non-technical way. We have assumed some familiarity with basic statistical ideas, but we have avoided detailed statistical theory. Instead, we have tried to write a practical guide that covers a wide range of methods. We emphasise the use of simple techniques in a variety of situations by using numerous examples, taken both from the literature and from our own experience. A number of these inevitably arise from our own particular field of interest - cancer clinical trials. This is also perhaps justifiable in that much of the pioneering work on quality of life assessment occurred in cancer, and cancer still remains the disease area that is associated with the largest number of quality of life instruments and the most publications. However, the issues that arise are common to quality of life assessment in general.

Acknowledgements

We would like to say a general thank you to all those with whom we have worked on aspects of quality of life over the years; especially, past and present members of the EORTC Quality of Life Study Group, and colleagues from the former MRC Cancer Therapy Committee Working Parties. Particular thanks go to Stein Kaasa of the Norwegian University of Science and Technology at Trondheim who permitted PMF to work on this book whilst on sabbatical and whose ideas greatly influenced our thinking about quality of life, and to Kristin Bjordal of The Radium Hospital, Oslo, who made extensive input and comments on many chapters and provided quality of life data that we used in examples. Finn Wisløff, for the Nordic Myeloma Study Group, very kindly allowed us to make extensive use their QoL data for many examples. We are grateful to the National Medical Research Council of Singapore for providing funds and facilities to enable us to complete this work. We also thank Dr Julian Thumboo, Tan Tock Seng Hospital, Singapore, for valuable comments on several chapters. Several chapters, and Chapter 7 in particular, were strongly influenced by manuals and guidelines published by the EORTC Quality of Life Study Group.

Peter M. Fayers and David Machin
January 2000

List of abbreviations

Note: We have adopted the policy of using italics to indicate variables, or things that take values.

α

Alpha, Type I error

αCronbach

Cronbach’s reliability coefficient

β

Beta, Power, 1-Type II error

κ

Kappa, Cohen’s measure of agreement

θ

Theta, an unobservable or “latent” variable

ρ

Rho, the correlation coefficient

σ

Sigma, the population standard deviation, estimated by SD

ADL

Activities of daily living

ANOVA

Analysis of variance

ARR

Absolute risk reduction

AUC

Area under the curve

CAT

Computer-adaptive test

CFA

Confirmatory factor analysis

CI

Confidence interval

CONSORT

Consolidated Standards of Reporting Trials

(http://www.consort-statement.org/)

CPMP

Committee for Proprietary Medicinal Products (European regulatory body)

DCE

Discrete choice experiment

df

Degrees of freedom

DIF

Differential item functioning

EF

Emotional functioning

EFA

Exploratory factor analysis

ES

Effect size

F-statistic

The ratio of two variance estimates; also called F-ratio

F-test

The statistical test used in ANOVA, based on the F-statistic

GEE

Generalised estimating equation

HYE

Healthy-years equivalent

IADL

Instrumental activities of daily living

ICC

Intraclass correlation

ICC

Item characteristic curve

IRT

Item response theory

LASA

Linear analogue self-assessment (scale)

MANOVA

Multivariate analysis of variance

MAR

Missing at random

MCAR

Missing completely at random

MIMIC

Multiple indicator multiple cause (model)

ML

Maximum likelihood (estimation)

MNAR

Missing not at random

MTMM

Multitrait–multimethod

NNT

Number needed to treat

NS

Not statistically significant

OR

Odds ratio

p

Probability, as in p-value

PF

Physical functioning

PRO

Patient-reported outcome

QALY

Quality-adjusted life years

QoL

Quality of life

Q-TWiST

Quality-adjusted time without symptoms and toxicity

RCT

Randomised controlled trial

RE

Relative efficiency

RR

Relative risk

RV

Relative validity

SD

Standard deviation of a sample

SE

Standard error

SEM

Standard error of measurement

SEM

Structured equation model

SG

Standard gamble

SMD

Standardised mean difference

SRM

Standardised response mean

t

Student’s t-statistic

TTO

Time trade-off

TWiST

Time without symptoms and toxicity

VAS

Visual analogue scale

WMD

Weighted mean difference

WTP

Willingness to pay

QoL instruments

 

AIMS

Arthritis Impact Measurement Scale

AMPS

Assessment of Motor and Process Skills

AQLQ

Asthma Quality of Life Questionnaire

BDI

Beck Depression Inventory

BI

Barthel Index of disability

BPI

Brief Pain Inventory

BPRS

Brief Psychiatric Rating Scale

EORTC QLQ-C30

European Organisation for Research and Treatment of

Cancer, Quality of Life Questionnaire, 30-items

EQ-5D

EuroQoL EQ-5D self report questionnaire

FACT-G

Functional Assessment of Cancer–General Version

FAQ

Functional Activity Questionnaire

FLIC

Functional Living Index–Cancer

GPH

General Perceived Health

HADS

Hospital Anxiety and Depression Scale

HDQoL Huntington’s Disease health-related Quality of Life

questionnaire

HOPES

HIV Overview of Problems–Evaluation System

HRSD

Hamilton Rating Scale for Depression

HUI

Health Utilities Index

MFI-20

Multidimensional Fatigue Inventory 20

MMSE

Mini-Mental State Examination

MPQ

McGill Pain Questionnaire

NHP

Nottingham Health Profile

PACIS

Perceived Adjustment to Chronic Illness Scale

PAQLQ

Pediatric Asthma Quality of Life Questionnaire

PASS

Pain Anxiety Symptoms Scale

PCQLI

Pediatric Cardiac Quality of Life Inventory

PGI

Patient Generated Index

POMS

Profile of Mood States

QOLIE-89

Quality of Life in Epilepsy

RSCL

Rotterdam Symptom Checklist

SEIQoL

Schedule for Evaluation of Individual Quality of Life

SF-36

Short Form 36

SIP

Sickness Impact Profile

WPSI

Washington Psychosocial Seizure Inventory

PART 1

Developing and Validating Instruments for Assessing Quality of Life and Patient-Reported Outcomes