Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with age-related macular degeneration

Nalin Payakachat, Kent H. Summers, Andreas M. Pleil, Matthew M. Murawski, Joseph Thomas, Kristofer Jennings, James G. Anderson

Research output: Contribution to journalArticle

35 Citations (Scopus)

Abstract

Purpose: In this study, we explored different statistical approaches to identify the best algorithm to predict EQ-5D utility scores from the NEI-VFQ 25 in patients with age-related macular degeneration (AMD). Methods: Ordinary least squares (OLS), Tobit, and censored least absolute deviation (CLAD) approaches were compared using cross-sectional data (primary dataset, n = 151) at screening from a phase I/II clinical trial in patients with AMD. Three models were specified in this study: full (includes all 12 dimensions of the NEI-VFQ 25), short (includes only the general health dimension and the composite score), and reduced model (using stepwise regression). To evaluate the predictive accuracy of the models, the mean absolute prediction error (MAPE), mean error, and root means squared error were calculated using in-sample cross-validation (within the primary dataset) and out-of-sample validation using an independent dataset (n = 393). The model that provided the lowest prediction errors was chosen as the best model. Results: In-sample cross-validation and out-of-sample validation consistently demonstrated that, compared to other approaches, heteroscedasticity-adjusted OLS produced the lowest MAPE (mean values were 0.1400, 0.1593, respectively) for the full model, while CLAD performed best for the short and reduced models (mean values were 0.1299, 0.1483, respectively). The normality and homoscedasticity assumptions of both OLS and Tobit were rejected. CLAD, however, can accommodate these particular violations. Conclusions: The CLAD-short model is recommended for producing the EQ-5D utility scores when only the NEI-VFQ 25 data are available.

Original languageEnglish (US)
Pages (from-to)801-813
Number of pages13
JournalQuality of Life Research
Volume18
Issue number7
DOIs
StatePublished - Sep 2009
Externally publishedYes

Fingerprint

National Eye Institute (U.S.)
Macular Degeneration
Least-Squares Analysis
Phase II Clinical Trials
Clinical Trials, Phase I
Health
Surveys and Questionnaires
Datasets

Keywords

  • Age-related macular degeneration (AMD)
  • EQ-5D
  • Low vision
  • Mapping
  • NEI-VFQ 25
  • Quality of life

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health

Cite this

Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with age-related macular degeneration. / Payakachat, Nalin; Summers, Kent H.; Pleil, Andreas M.; Murawski, Matthew M.; Thomas, Joseph; Jennings, Kristofer; Anderson, James G.

In: Quality of Life Research, Vol. 18, No. 7, 09.2009, p. 801-813.

Research output: Contribution to journalArticle

Payakachat, Nalin ; Summers, Kent H. ; Pleil, Andreas M. ; Murawski, Matthew M. ; Thomas, Joseph ; Jennings, Kristofer ; Anderson, James G. / Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with age-related macular degeneration. In: Quality of Life Research. 2009 ; Vol. 18, No. 7. pp. 801-813.
@article{5e1be10501004321a6f4753a62a3bc25,
title = "Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with age-related macular degeneration",
abstract = "Purpose: In this study, we explored different statistical approaches to identify the best algorithm to predict EQ-5D utility scores from the NEI-VFQ 25 in patients with age-related macular degeneration (AMD). Methods: Ordinary least squares (OLS), Tobit, and censored least absolute deviation (CLAD) approaches were compared using cross-sectional data (primary dataset, n = 151) at screening from a phase I/II clinical trial in patients with AMD. Three models were specified in this study: full (includes all 12 dimensions of the NEI-VFQ 25), short (includes only the general health dimension and the composite score), and reduced model (using stepwise regression). To evaluate the predictive accuracy of the models, the mean absolute prediction error (MAPE), mean error, and root means squared error were calculated using in-sample cross-validation (within the primary dataset) and out-of-sample validation using an independent dataset (n = 393). The model that provided the lowest prediction errors was chosen as the best model. Results: In-sample cross-validation and out-of-sample validation consistently demonstrated that, compared to other approaches, heteroscedasticity-adjusted OLS produced the lowest MAPE (mean values were 0.1400, 0.1593, respectively) for the full model, while CLAD performed best for the short and reduced models (mean values were 0.1299, 0.1483, respectively). The normality and homoscedasticity assumptions of both OLS and Tobit were rejected. CLAD, however, can accommodate these particular violations. Conclusions: The CLAD-short model is recommended for producing the EQ-5D utility scores when only the NEI-VFQ 25 data are available.",
keywords = "Age-related macular degeneration (AMD), EQ-5D, Low vision, Mapping, NEI-VFQ 25, Quality of life",
author = "Nalin Payakachat and Summers, {Kent H.} and Pleil, {Andreas M.} and Murawski, {Matthew M.} and Joseph Thomas and Kristofer Jennings and Anderson, {James G.}",
year = "2009",
month = "9",
doi = "10.1007/s11136-009-9499-6",
language = "English (US)",
volume = "18",
pages = "801--813",
journal = "Quality of Life Research",
issn = "0962-9343",
publisher = "Springer Netherlands",
number = "7",

}

TY - JOUR

T1 - Predicting EQ-5D utility scores from the 25-item National Eye Institute Vision Function Questionnaire (NEI-VFQ 25) in patients with age-related macular degeneration

AU - Payakachat, Nalin

AU - Summers, Kent H.

AU - Pleil, Andreas M.

AU - Murawski, Matthew M.

AU - Thomas, Joseph

AU - Jennings, Kristofer

AU - Anderson, James G.

PY - 2009/9

Y1 - 2009/9

N2 - Purpose: In this study, we explored different statistical approaches to identify the best algorithm to predict EQ-5D utility scores from the NEI-VFQ 25 in patients with age-related macular degeneration (AMD). Methods: Ordinary least squares (OLS), Tobit, and censored least absolute deviation (CLAD) approaches were compared using cross-sectional data (primary dataset, n = 151) at screening from a phase I/II clinical trial in patients with AMD. Three models were specified in this study: full (includes all 12 dimensions of the NEI-VFQ 25), short (includes only the general health dimension and the composite score), and reduced model (using stepwise regression). To evaluate the predictive accuracy of the models, the mean absolute prediction error (MAPE), mean error, and root means squared error were calculated using in-sample cross-validation (within the primary dataset) and out-of-sample validation using an independent dataset (n = 393). The model that provided the lowest prediction errors was chosen as the best model. Results: In-sample cross-validation and out-of-sample validation consistently demonstrated that, compared to other approaches, heteroscedasticity-adjusted OLS produced the lowest MAPE (mean values were 0.1400, 0.1593, respectively) for the full model, while CLAD performed best for the short and reduced models (mean values were 0.1299, 0.1483, respectively). The normality and homoscedasticity assumptions of both OLS and Tobit were rejected. CLAD, however, can accommodate these particular violations. Conclusions: The CLAD-short model is recommended for producing the EQ-5D utility scores when only the NEI-VFQ 25 data are available.

AB - Purpose: In this study, we explored different statistical approaches to identify the best algorithm to predict EQ-5D utility scores from the NEI-VFQ 25 in patients with age-related macular degeneration (AMD). Methods: Ordinary least squares (OLS), Tobit, and censored least absolute deviation (CLAD) approaches were compared using cross-sectional data (primary dataset, n = 151) at screening from a phase I/II clinical trial in patients with AMD. Three models were specified in this study: full (includes all 12 dimensions of the NEI-VFQ 25), short (includes only the general health dimension and the composite score), and reduced model (using stepwise regression). To evaluate the predictive accuracy of the models, the mean absolute prediction error (MAPE), mean error, and root means squared error were calculated using in-sample cross-validation (within the primary dataset) and out-of-sample validation using an independent dataset (n = 393). The model that provided the lowest prediction errors was chosen as the best model. Results: In-sample cross-validation and out-of-sample validation consistently demonstrated that, compared to other approaches, heteroscedasticity-adjusted OLS produced the lowest MAPE (mean values were 0.1400, 0.1593, respectively) for the full model, while CLAD performed best for the short and reduced models (mean values were 0.1299, 0.1483, respectively). The normality and homoscedasticity assumptions of both OLS and Tobit were rejected. CLAD, however, can accommodate these particular violations. Conclusions: The CLAD-short model is recommended for producing the EQ-5D utility scores when only the NEI-VFQ 25 data are available.

KW - Age-related macular degeneration (AMD)

KW - EQ-5D

KW - Low vision

KW - Mapping

KW - NEI-VFQ 25

KW - Quality of life

UR - http://www.scopus.com/inward/record.url?scp=69249211023&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=69249211023&partnerID=8YFLogxK

U2 - 10.1007/s11136-009-9499-6

DO - 10.1007/s11136-009-9499-6

M3 - Article

VL - 18

SP - 801

EP - 813

JO - Quality of Life Research

JF - Quality of Life Research

SN - 0962-9343

IS - 7

ER -