Peter Fayers, Emeritus Professor of Medical Statistics, University of Aberdeen, UK; Norwegian University of Science and Technology (NTNU), Trondheim, Norway. David Machin, Emeritus Professor of Clinical Trials Research, University of Sheffield, UK and Emeritus Professor of Clinical Statistics, University of Leicester.

Select Guide Rating

• Preface to the third edition xiiiPreface to the second edition xvPreface to the first edition xviiList of abbreviations xixPart 1 Developing and Validating Instruments for AssessingQuality of Life and Patient-Reported Outcomes1 Introduction 31.1 Patient‐reported outcomes 31.2 What is a patient‐reported outcome? 41.3 What is quality of life? 41.4 Historical development 61.5 Why measure quality of life? 91.6 Which clinical trials should assess QoL? 171.7 How to measure quality of life 181.8 Instruments 191.9 Computer‐adaptive instruments 321.10 Conclusions 322 Principles of measurement scales 352.1 Introduction 352.2 Scales and items 352.3 Constructs and latent variables 362.4 Single global questions versus multi‐item scales 372.5 Single‐item versus multi‐item scales 402.6 Effect indicators and causal indicators 422.7 Psychometrics, factor analysis and item response theory 482.8 Psychometric versus clinimetric scales 522.9 Sufficient causes, necessary causes and scoring items 532.10 Discriminative, evaluative and predictive instruments 542.11 Measuring quality of life: reflective, causal and composite indicators? 552.12 Further reading 562.13 Conclusions 563 Developing a questionnaire 573.1 Introduction 573.2 General issues 583.3 Defining the target population 583.4 Phases of development 593.5 Phase 1: Generation of issues 613.6 Qualitative methods 633.7 Sample sizes 663.8 Phase 2: Developing items 683.9 Multi‐item scales 723.10 Wording of questions 733.11 Face and content validity of the proposed questionnaire 743.12 Phase 3: Pre‐testing the questionnaire 743.13 Cognitive interviewing 773.14 Translation 803.15 Phase 4: Field‐testing 803.16 Conclusions 863.17 Further reading 874 Scores and measurements: validity, reliability, sensitivity 894.1 Introduction 894.2 Content validity 904.3 Criterion validity 944.4 Construct validity 964.5 Repeated assessments and change over time 1044.6 Reliability 1044.7 Sensitivity and responsiveness 1174.8 Conclusions 1244.9 Further reading 1245 Multi‐item scales 1255.1 Introduction 1255.2 Significance tests 1265.3 Correlations 1275.4 Construct validity 1335.5 Cronbach’s α and internal consistency 1395.6 Validation or alteration? 1435.7 Implications for formative or causal items 1445.8 Conclusions 1476 Factor analysis and structural equation modelling 1496.1 Introduction 1496.2 Correlation patterns 1506.3 Path diagrams 1526.4 Factor analysis 1546.5 Factor analysis of the HADS questionnaire 1546.6 Uses of factor analysis 1596.7 Applying factor analysis: Choices and decisions 1616.8 Assumptions for factor analysis 1676.9 Factor analysis in QoL research 1716.10 Limitations of correlation-based analysis 1726.11 Formative or causal models 1736.12 Confirmatory factor analysis and structural equation modelling 1766.13 Chi-square goodness-of-fit test 1786.14 Approximate goodness-of-fit indices 1806.15 Comparative fit of models 1816.16 Difficulty-factors 1826.17 Bifactor analysis 1836.18 Do formative or causal relationships matter? 1866.19 Conclusions 1876.20 Further reading, and software 1887 Item response theory and differential item functioning 1897.1 Introduction 1897.2 Item characteristic curves 1917.3 Logistic models 1937.4 Polytomous item response theory models 1967.5 Applying logistic IRT models 1977.6 Assumptions of IRT models 2057.7 Fitting item response theory models: Tips 2087.8 Test design and validation 2097.9 IRT versus traditional and Guttman scales 2097.10 Differential item functioning 2107.11 Sample size for DIF analyses 2187.12 Quantifying differential item functioning 2197.13 Exploring differential item functioning: Tips 2197.14 Conclusions 2217.15 Further reading, and software 2228 Item banks, item linking and computer-adaptive tests 2238.1 Introduction 2238.2 Item bank 2248.3 Item evaluation, reduction and calibration 2268.4 Item linking and test equating 2288.5 Test information 2318.6 Computer-adaptive testing 2328.7 Stopping rules and simulations 2358.8 Computer-adaptive testing software 2368.9 CATs for PROs 2378.10 Computer-assisted tests 2388.11 Short-form tests 2398.12 Conclusions 2398.13 Further reading 240Part 2 Assessing, Analysing and Reporting Patient-Reported Outcomes and the Quality of Life of Patients9 Choosing and scoring questionnaires 2439.1 Introduction 2439.2 Finding instruments 2449.3 Generic versus specific 2459.4 Content and presentation 2469.5 Choice of instrument 2479.6 Scoring multi-item scales 2509.7 Conclusions 2569.8 Further reading 25710 Clinical trials 25910.1 Introduction 25910.2 Basic design issues 26010.3 Compliance 26210.4 Administering a quality‐of‐life assessment 26810.5 Recommendations for writing protocols 27010.6 Standard operating procedures 28010.7 Summary and checklist 28110.8 Further reading 28211 Sample sizes 28311.1 Introduction 28311.2 Significance tests, p‐values and power 28411.3 Estimating sample size 28411.4 Comparing two groups 28911.5 Comparison with a reference population 29811.6 Non‐inferiority studies 29811.7 Choice of sample size method 30111.8 Non‐Normal distributions 30211.9 Multiple testing 30311.10 Specifying the target difference 30511.11 Sample size estimation is pre‐study 30511.12 Attrition 30611.13 Circumspection 30611.14 Conclusion 30611.15 Further reading 30712 Cross‐sectional analysis 30912.1 Types of data 30912.2 Comparing two groups 31212.3 Adjusting for covariates 32412.4 Changes from baseline 33012.5 Analysis of variance 33112.6 Analysis of variance models 33612.7 Graphical summaries 33712.8 Endpoints 34212.9 Conclusions 34313 Exploring longitudinal data 34513.1 Area under the curve 34513.2 Graphical presentations 34813.3 Tabular presentations 35813.4 Reporting 36013.5 Conclusions 36514 Modelling longitudinal data 36714.1 Preliminaries 36714.2 Auto-correlation 36814.3 Repeated measures 37314.4 Other situations 38814.5 Modelling versus area under the curve 38914.6 Conclusions 39015 Missing data 39315.1 Introduction 39315.2 Why do missing data matter? 39615.3 Types of missing data 40015.4 Missing items 40315.5 Methods for missing items within a form 40415.6 Missing forms 40815.7 Methods for missing forms 41015.8 Simple methods for missing forms 41015.9 Methods of imputation that incorporate variability 41515.10 Multiple imputation 42115.11 Pattern mixture models 42215.12 Comments 42415.13 Degrees of freedom 42515.14 Sensitivity analysis 42615.15 Conclusions 42615.16 Further reading 42716 Practical and reporting issues 42916.1 Introduction 42916.2 The reporting of design issues 43016.3 Data analysis 43016.4 Elements of good graphics 43616.5 Some errors 44016.6 Guidelines for reporting 44216.7 Further reading 44517 Death, and quality-adjusted survival 44717.1 Introduction 44717.2 Attrition due to death 44817.3 Preferences and utilities 44917.4 Multi-attribute utility (MAU) measures 45317.5 Utility-based instruments 45417.6 Quality-adjusted life years (QALYs) 45617.7 Utilities for traditional instruments 45717.8 Q-TWiST 46217.9 Sensitivity analysis 46717.10 Prognosis and variation with time 47017.11 Alternatives to QALY 47217.12 Conclusions 47317.13 Further reading 47418 Clinical interpretation 47518.1 Introduction 47518.2 Statistical significance 47618.3 Absolute levels and changes over time 47718.4 Threshold values: percentages 47818.5 Population norms 47918.6 Minimal important difference 48818.7 Anchoring against other measurements 49218.8 Minimum detectable change 49318.9 Expert judgement for evidence-based guidelines 49418.10 Impact of the state of quality of life 49518.11 Changes in relation to life events 49618.12 Effect size statistics 49818.13 Patient variability 50518.14 Number needed to treat 50618.15 Conclusions 50918.16 Further reading 50919 Biased reporting and response shift 51119.1 Bias 51119.2 Recall bias 51219.3 Selective reporting bias 51319.4 Other biases affecting PROs 51419.5 Response shift 51619.6 Assessing response shift 52119.7 Impact of response shift 52319.8 Clinical trials 52319.9 Non‐randomised studies 52519.10 Conclusions 52620 Meta‐analysis 52720.1 Introduction 52720.2 Defining objectives 52820.3 Defining outcomes 52820.4 Literature searching 52820.5 Assessing quality 52920.6 Summarising results 53320.7 Measures of treatment effect 53420.8 Combining studies 53720.9 Forest plot 54220.10 Heterogeneity 54220.11 Publication bias and funnel plots 54420.12 Conclusions 54520.13 Further reading 546Appendix 1: Examples of instruments 547Generic instrumentsE1 Sickness Impact Profile (SIP) 549E2 Nottingham Health Profile (NHP) 551E3 SF36v2TM Health Survey Standard Version 552E4 EuroQoL EQ-5D-5L 555E5 Patient Generated Index of quality of life (PGI) 557Disease-specific instruments 559E6 European Organisation for Research and Treatment of Cancer QLQ-C30 (EORTC QLQ-C30) 559E7 Elderly cancer patients module (EORTC QLQ-ELD14) 561E8 Functional Assessment of Cancer Therapy – General (FACT-G) 562E9 Rotterdam Symptom Checklist (RSCL) 564E10 Quality of Life in Epilepsy Inventory (QOLIE-89) 566E11 Paediatric Asthma Quality of Life Questionnaire (PAQLQ) 570Domain-specific instruments 573E12 Hospital Anxiety and Depression Scale (HADS) 573E13 Short-Form McGill Pain Questionnaire (SF-MPQ) 574E14 Multidimensional Fatigue Inventory (MFI-20) 575ADL and disability 577E 15 (Modified) Barthel Index of Disability (MBI) 577Appendix 2: Statistical tables 579Table T1: Normal distribution 579Table T2: Probability points of the Normal distribution 581Table T3: Student’s t‐distribution 582Table T4: The χ2 distribution 583Table T5: The F‐distribution 584References 585Index 613

Peter Fayers, Emeritus Professor of Medical Statistics, University of Aberdeen, UK; Norwegian University of Science and Technology (NTNU), Trondheim, Norway. David Machin, Emeritus Professor of Clinical Trials Research, University of Sheffield, UK and Emeritus Professor of Clinical Statistics, University of Leicester.

Select Guide Rating

• Preface to the third edition xiiiPreface to the second edition xvPreface to the first edition xviiList of abbreviations xixPart 1 Developing and Validating Instruments for AssessingQuality of Life and Patient-Reported Outcomes1 Introduction 31.1 Patient‐reported outcomes 31.2 What is a patient‐reported outcome? 41.3 What is quality of life? 41.4 Historical development 61.5 Why measure quality of life? 91.6 Which clinical trials should assess QoL? 171.7 How to measure quality of life 181.8 Instruments 191.9 Computer‐adaptive instruments 321.10 Conclusions 322 Principles of measurement scales 352.1 Introduction 352.2 Scales and items 352.3 Constructs and latent variables 362.4 Single global questions versus multi‐item scales 372.5 Single‐item versus multi‐item scales 402.6 Effect indicators and causal indicators 422.7 Psychometrics, factor analysis and item response theory 482.8 Psychometric versus clinimetric scales 522.9 Sufficient causes, necessary causes and scoring items 532.10 Discriminative, evaluative and predictive instruments 542.11 Measuring quality of life: reflective, causal and composite indicators? 552.12 Further reading 562.13 Conclusions 563 Developing a questionnaire 573.1 Introduction 573.2 General issues 583.3 Defining the target population 583.4 Phases of development 593.5 Phase 1: Generation of issues 613.6 Qualitative methods 633.7 Sample sizes 663.8 Phase 2: Developing items 683.9 Multi‐item scales 723.10 Wording of questions 733.11 Face and content validity of the proposed questionnaire 743.12 Phase 3: Pre‐testing the questionnaire 743.13 Cognitive interviewing 773.14 Translation 803.15 Phase 4: Field‐testing 803.16 Conclusions 863.17 Further reading 874 Scores and measurements: validity, reliability, sensitivity 894.1 Introduction 894.2 Content validity 904.3 Criterion validity 944.4 Construct validity 964.5 Repeated assessments and change over time 1044.6 Reliability 1044.7 Sensitivity and responsiveness 1174.8 Conclusions 1244.9 Further reading 1245 Multi‐item scales 1255.1 Introduction 1255.2 Significance tests 1265.3 Correlations 1275.4 Construct validity 1335.5 Cronbach’s α and internal consistency 1395.6 Validation or alteration? 1435.7 Implications for formative or causal items 1445.8 Conclusions 1476 Factor analysis and structural equation modelling 1496.1 Introduction 1496.2 Correlation patterns 1506.3 Path diagrams 1526.4 Factor analysis 1546.5 Factor analysis of the HADS questionnaire 1546.6 Uses of factor analysis 1596.7 Applying factor analysis: Choices and decisions 1616.8 Assumptions for factor analysis 1676.9 Factor analysis in QoL research 1716.10 Limitations of correlation-based analysis 1726.11 Formative or causal models 1736.12 Confirmatory factor analysis and structural equation modelling 1766.13 Chi-square goodness-of-fit test 1786.14 Approximate goodness-of-fit indices 1806.15 Comparative fit of models 1816.16 Difficulty-factors 1826.17 Bifactor analysis 1836.18 Do formative or causal relationships matter? 1866.19 Conclusions 1876.20 Further reading, and software 1887 Item response theory and differential item functioning 1897.1 Introduction 1897.2 Item characteristic curves 1917.3 Logistic models 1937.4 Polytomous item response theory models 1967.5 Applying logistic IRT models 1977.6 Assumptions of IRT models 2057.7 Fitting item response theory models: Tips 2087.8 Test design and validation 2097.9 IRT versus traditional and Guttman scales 2097.10 Differential item functioning 2107.11 Sample size for DIF analyses 2187.12 Quantifying differential item functioning 2197.13 Exploring differential item functioning: Tips 2197.14 Conclusions 2217.15 Further reading, and software 2228 Item banks, item linking and computer-adaptive tests 2238.1 Introduction 2238.2 Item bank 2248.3 Item evaluation, reduction and calibration 2268.4 Item linking and test equating 2288.5 Test information 2318.6 Computer-adaptive testing 2328.7 Stopping rules and simulations 2358.8 Computer-adaptive testing software 2368.9 CATs for PROs 2378.10 Computer-assisted tests 2388.11 Short-form tests 2398.12 Conclusions 2398.13 Further reading 240Part 2 Assessing, Analysing and Reporting Patient-Reported Outcomes and the Quality of Life of Patients9 Choosing and scoring questionnaires 2439.1 Introduction 2439.2 Finding instruments 2449.3 Generic versus specific 2459.4 Content and presentation 2469.5 Choice of instrument 2479.6 Scoring multi-item scales 2509.7 Conclusions 2569.8 Further reading 25710 Clinical trials 25910.1 Introduction 25910.2 Basic design issues 26010.3 Compliance 26210.4 Administering a quality‐of‐life assessment 26810.5 Recommendations for writing protocols 27010.6 Standard operating procedures 28010.7 Summary and checklist 28110.8 Further reading 28211 Sample sizes 28311.1 Introduction 28311.2 Significance tests, p‐values and power 28411.3 Estimating sample size 28411.4 Comparing two groups 28911.5 Comparison with a reference population 29811.6 Non‐inferiority studies 29811.7 Choice of sample size method 30111.8 Non‐Normal distributions 30211.9 Multiple testing 30311.10 Specifying the target difference 30511.11 Sample size estimation is pre‐study 30511.12 Attrition 30611.13 Circumspection 30611.14 Conclusion 30611.15 Further reading 30712 Cross‐sectional analysis 30912.1 Types of data 30912.2 Comparing two groups 31212.3 Adjusting for covariates 32412.4 Changes from baseline 33012.5 Analysis of variance 33112.6 Analysis of variance models 33612.7 Graphical summaries 33712.8 Endpoints 34212.9 Conclusions 34313 Exploring longitudinal data 34513.1 Area under the curve 34513.2 Graphical presentations 34813.3 Tabular presentations 35813.4 Reporting 36013.5 Conclusions 36514 Modelling longitudinal data 36714.1 Preliminaries 36714.2 Auto-correlation 36814.3 Repeated measures 37314.4 Other situations 38814.5 Modelling versus area under the curve 38914.6 Conclusions 39015 Missing data 39315.1 Introduction 39315.2 Why do missing data matter? 39615.3 Types of missing data 40015.4 Missing items 40315.5 Methods for missing items within a form 40415.6 Missing forms 40815.7 Methods for missing forms 41015.8 Simple methods for missing forms 41015.9 Methods of imputation that incorporate variability 41515.10 Multiple imputation 42115.11 Pattern mixture models 42215.12 Comments 42415.13 Degrees of freedom 42515.14 Sensitivity analysis 42615.15 Conclusions 42615.16 Further reading 42716 Practical and reporting issues 42916.1 Introduction 42916.2 The reporting of design issues 43016.3 Data analysis 43016.4 Elements of good graphics 43616.5 Some errors 44016.6 Guidelines for reporting 44216.7 Further reading 44517 Death, and quality-adjusted survival 44717.1 Introduction 44717.2 Attrition due to death 44817.3 Preferences and utilities 44917.4 Multi-attribute utility (MAU) measures 45317.5 Utility-based instruments 45417.6 Quality-adjusted life years (QALYs) 45617.7 Utilities for traditional instruments 45717.8 Q-TWiST 46217.9 Sensitivity analysis 46717.10 Prognosis and variation with time 47017.11 Alternatives to QALY 47217.12 Conclusions 47317.13 Further reading 47418 Clinical interpretation 47518.1 Introduction 47518.2 Statistical significance 47618.3 Absolute levels and changes over time 47718.4 Threshold values: percentages 47818.5 Population norms 47918.6 Minimal important difference 48818.7 Anchoring against other measurements 49218.8 Minimum detectable change 49318.9 Expert judgement for evidence-based guidelines 49418.10 Impact of the state of quality of life 49518.11 Changes in relation to life events 49618.12 Effect size statistics 49818.13 Patient variability 50518.14 Number needed to treat 50618.15 Conclusions 50918.16 Further reading 50919 Biased reporting and response shift 51119.1 Bias 51119.2 Recall bias 51219.3 Selective reporting bias 51319.4 Other biases affecting PROs 51419.5 Response shift 51619.6 Assessing response shift 52119.7 Impact of response shift 52319.8 Clinical trials 52319.9 Non‐randomised studies 52519.10 Conclusions 52620 Meta‐analysis 52720.1 Introduction 52720.2 Defining objectives 52820.3 Defining outcomes 52820.4 Literature searching 52820.5 Assessing quality 52920.6 Summarising results 53320.7 Measures of treatment effect 53420.8 Combining studies 53720.9 Forest plot 54220.10 Heterogeneity 54220.11 Publication bias and funnel plots 54420.12 Conclusions 54520.13 Further reading 546Appendix 1: Examples of instruments 547Generic instrumentsE1 Sickness Impact Profile (SIP) 549E2 Nottingham Health Profile (NHP) 551E3 SF36v2TM Health Survey Standard Version 552E4 EuroQoL EQ-5D-5L 555E5 Patient Generated Index of quality of life (PGI) 557Disease-specific instruments 559E6 European Organisation for Research and Treatment of Cancer QLQ-C30 (EORTC QLQ-C30) 559E7 Elderly cancer patients module (EORTC QLQ-ELD14) 561E8 Functional Assessment of Cancer Therapy – General (FACT-G) 562E9 Rotterdam Symptom Checklist (RSCL) 564E10 Quality of Life in Epilepsy Inventory (QOLIE-89) 566E11 Paediatric Asthma Quality of Life Questionnaire (PAQLQ) 570Domain-specific instruments 573E12 Hospital Anxiety and Depression Scale (HADS) 573E13 Short-Form McGill Pain Questionnaire (SF-MPQ) 574E14 Multidimensional Fatigue Inventory (MFI-20) 575ADL and disability 577E 15 (Modified) Barthel Index of Disability (MBI) 577Appendix 2: Statistical tables 579Table T1: Normal distribution 579Table T2: Probability points of the Normal distribution 581Table T3: Student’s t‐distribution 582Table T4: The χ2 distribution 583Table T5: The F‐distribution 584References 585Index 613

Peter Fayers, Emeritus Professor of Medical Statistics, University of Aberdeen, UK; Norwegian University of Science and Technology (NTNU), Trondheim, Norway. David Machin, Emeritus Professor of Clinical Trials Research, University of Sheffield, UK and Emeritus Professor of Clinical Statistics, University of Leicester.

Select Guide Rating

• Preface to the third edition xiiiPreface to the second edition xvPreface to the first edition xviiList of abbreviations xixPart 1 Developing and Validating Instruments for AssessingQuality of Life and Patient-Reported Outcomes1 Introduction 31.1 Patient‐reported outcomes 31.2 What is a patient‐reported outcome? 41.3 What is quality of life? 41.4 Historical development 61.5 Why measure quality of life? 91.6 Which clinical trials should assess QoL? 171.7 How to measure quality of life 181.8 Instruments 191.9 Computer‐adaptive instruments 321.10 Conclusions 322 Principles of measurement scales 352.1 Introduction 352.2 Scales and items 352.3 Constructs and latent variables 362.4 Single global questions versus multi‐item scales 372.5 Single‐item versus multi‐item scales 402.6 Effect indicators and causal indicators 422.7 Psychometrics, factor analysis and item response theory 482.8 Psychometric versus clinimetric scales 522.9 Sufficient causes, necessary causes and scoring items 532.10 Discriminative, evaluative and predictive instruments 542.11 Measuring quality of life: reflective, causal and composite indicators? 552.12 Further reading 562.13 Conclusions 563 Developing a questionnaire 573.1 Introduction 573.2 General issues 583.3 Defining the target population 583.4 Phases of development 593.5 Phase 1: Generation of issues 613.6 Qualitative methods 633.7 Sample sizes 663.8 Phase 2: Developing items 683.9 Multi‐item scales 723.10 Wording of questions 733.11 Face and content validity of the proposed questionnaire 743.12 Phase 3: Pre‐testing the questionnaire 743.13 Cognitive interviewing 773.14 Translation 803.15 Phase 4: Field‐testing 803.16 Conclusions 863.17 Further reading 874 Scores and measurements: validity, reliability, sensitivity 894.1 Introduction 894.2 Content validity 904.3 Criterion validity 944.4 Construct validity 964.5 Repeated assessments and change over time 1044.6 Reliability 1044.7 Sensitivity and responsiveness 1174.8 Conclusions 1244.9 Further reading 1245 Multi‐item scales 1255.1 Introduction 1255.2 Significance tests 1265.3 Correlations 1275.4 Construct validity 1335.5 Cronbach’s α and internal consistency 1395.6 Validation or alteration? 1435.7 Implications for formative or causal items 1445.8 Conclusions 1476 Factor analysis and structural equation modelling 1496.1 Introduction 1496.2 Correlation patterns 1506.3 Path diagrams 1526.4 Factor analysis 1546.5 Factor analysis of the HADS questionnaire 1546.6 Uses of factor analysis 1596.7 Applying factor analysis: Choices and decisions 1616.8 Assumptions for factor analysis 1676.9 Factor analysis in QoL research 1716.10 Limitations of correlation-based analysis 1726.11 Formative or causal models 1736.12 Confirmatory factor analysis and structural equation modelling 1766.13 Chi-square goodness-of-fit test 1786.14 Approximate goodness-of-fit indices 1806.15 Comparative fit of models 1816.16 Difficulty-factors 1826.17 Bifactor analysis 1836.18 Do formative or causal relationships matter? 1866.19 Conclusions 1876.20 Further reading, and software 1887 Item response theory and differential item functioning 1897.1 Introduction 1897.2 Item characteristic curves 1917.3 Logistic models 1937.4 Polytomous item response theory models 1967.5 Applying logistic IRT models 1977.6 Assumptions of IRT models 2057.7 Fitting item response theory models: Tips 2087.8 Test design and validation 2097.9 IRT versus traditional and Guttman scales 2097.10 Differential item functioning 2107.11 Sample size for DIF analyses 2187.12 Quantifying differential item functioning 2197.13 Exploring differential item functioning: Tips 2197.14 Conclusions 2217.15 Further reading, and software 2228 Item banks, item linking and computer-adaptive tests 2238.1 Introduction 2238.2 Item bank 2248.3 Item evaluation, reduction and calibration 2268.4 Item linking and test equating 2288.5 Test information 2318.6 Computer-adaptive testing 2328.7 Stopping rules and simulations 2358.8 Computer-adaptive testing software 2368.9 CATs for PROs 2378.10 Computer-assisted tests 2388.11 Short-form tests 2398.12 Conclusions 2398.13 Further reading 240Part 2 Assessing, Analysing and Reporting Patient-Reported Outcomes and the Quality of Life of Patients9 Choosing and scoring questionnaires 2439.1 Introduction 2439.2 Finding instruments 2449.3 Generic versus specific 2459.4 Content and presentation 2469.5 Choice of instrument 2479.6 Scoring multi-item scales 2509.7 Conclusions 2569.8 Further reading 25710 Clinical trials 25910.1 Introduction 25910.2 Basic design issues 26010.3 Compliance 26210.4 Administering a quality‐of‐life assessment 26810.5 Recommendations for writing protocols 27010.6 Standard operating procedures 28010.7 Summary and checklist 28110.8 Further reading 28211 Sample sizes 28311.1 Introduction 28311.2 Significance tests, p‐values and power 28411.3 Estimating sample size 28411.4 Comparing two groups 28911.5 Comparison with a reference population 29811.6 Non‐inferiority studies 29811.7 Choice of sample size method 30111.8 Non‐Normal distributions 30211.9 Multiple testing 30311.10 Specifying the target difference 30511.11 Sample size estimation is pre‐study 30511.12 Attrition 30611.13 Circumspection 30611.14 Conclusion 30611.15 Further reading 30712 Cross‐sectional analysis 30912.1 Types of data 30912.2 Comparing two groups 31212.3 Adjusting for covariates 32412.4 Changes from baseline 33012.5 Analysis of variance 33112.6 Analysis of variance models 33612.7 Graphical summaries 33712.8 Endpoints 34212.9 Conclusions 34313 Exploring longitudinal data 34513.1 Area under the curve 34513.2 Graphical presentations 34813.3 Tabular presentations 35813.4 Reporting 36013.5 Conclusions 36514 Modelling longitudinal data 36714.1 Preliminaries 36714.2 Auto-correlation 36814.3 Repeated measures 37314.4 Other situations 38814.5 Modelling versus area under the curve 38914.6 Conclusions 39015 Missing data 39315.1 Introduction 39315.2 Why do missing data matter? 39615.3 Types of missing data 40015.4 Missing items 40315.5 Methods for missing items within a form 40415.6 Missing forms 40815.7 Methods for missing forms 41015.8 Simple methods for missing forms 41015.9 Methods of imputation that incorporate variability 41515.10 Multiple imputation 42115.11 Pattern mixture models 42215.12 Comments 42415.13 Degrees of freedom 42515.14 Sensitivity analysis 42615.15 Conclusions 42615.16 Further reading 42716 Practical and reporting issues 42916.1 Introduction 42916.2 The reporting of design issues 43016.3 Data analysis 43016.4 Elements of good graphics 43616.5 Some errors 44016.6 Guidelines for reporting 44216.7 Further reading 44517 Death, and quality-adjusted survival 44717.1 Introduction 44717.2 Attrition due to death 44817.3 Preferences and utilities 44917.4 Multi-attribute utility (MAU) measures 45317.5 Utility-based instruments 45417.6 Quality-adjusted life years (QALYs) 45617.7 Utilities for traditional instruments 45717.8 Q-TWiST 46217.9 Sensitivity analysis 46717.10 Prognosis and variation with time 47017.11 Alternatives to QALY 47217.12 Conclusions 47317.13 Further reading 47418 Clinical interpretation 47518.1 Introduction 47518.2 Statistical significance 47618.3 Absolute levels and changes over time 47718.4 Threshold values: percentages 47818.5 Population norms 47918.6 Minimal important difference 48818.7 Anchoring against other measurements 49218.8 Minimum detectable change 49318.9 Expert judgement for evidence-based guidelines 49418.10 Impact of the state of quality of life 49518.11 Changes in relation to life events 49618.12 Effect size statistics 49818.13 Patient variability 50518.14 Number needed to treat 50618.15 Conclusions 50918.16 Further reading 50919 Biased reporting and response shift 51119.1 Bias 51119.2 Recall bias 51219.3 Selective reporting bias 51319.4 Other biases affecting PROs 51419.5 Response shift 51619.6 Assessing response shift 52119.7 Impact of response shift 52319.8 Clinical trials 52319.9 Non‐randomised studies 52519.10 Conclusions 52620 Meta‐analysis 52720.1 Introduction 52720.2 Defining objectives 52820.3 Defining outcomes 52820.4 Literature searching 52820.5 Assessing quality 52920.6 Summarising results 53320.7 Measures of treatment effect 53420.8 Combining studies 53720.9 Forest plot 54220.10 Heterogeneity 54220.11 Publication bias and funnel plots 54420.12 Conclusions 54520.13 Further reading 546Appendix 1: Examples of instruments 547Generic instrumentsE1 Sickness Impact Profile (SIP) 549E2 Nottingham Health Profile (NHP) 551E3 SF36v2TM Health Survey Standard Version 552E4 EuroQoL EQ-5D-5L 555E5 Patient Generated Index of quality of life (PGI) 557Disease-specific instruments 559E6 European Organisation for Research and Treatment of Cancer QLQ-C30 (EORTC QLQ-C30) 559E7 Elderly cancer patients module (EORTC QLQ-ELD14) 561E8 Functional Assessment of Cancer Therapy – General (FACT-G) 562E9 Rotterdam Symptom Checklist (RSCL) 564E10 Quality of Life in Epilepsy Inventory (QOLIE-89) 566E11 Paediatric Asthma Quality of Life Questionnaire (PAQLQ) 570Domain-specific instruments 573E12 Hospital Anxiety and Depression Scale (HADS) 573E13 Short-Form McGill Pain Questionnaire (SF-MPQ) 574E14 Multidimensional Fatigue Inventory (MFI-20) 575ADL and disability 577E 15 (Modified) Barthel Index of Disability (MBI) 577Appendix 2: Statistical tables 579Table T1: Normal distribution 579Table T2: Probability points of the Normal distribution 581Table T3: Student’s t‐distribution 582Table T4: The χ2 distribution 583Table T5: The F‐distribution 584References 585Index 613