Exploring the Benefits of Transformations in Health Utility Mapping

Mitsakakis, Nicholas; Bremner, Karen E.; Tomlinson, George; Krahn, Murray

doi:10.1177/0272989x19896567

Cited by 3 publications

(5 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As iHOT-33 and HAGOS scores are anchored by values of 0 and 100, they may not always be optimally modeled using linear regression. To validate the results of our main analysis, we conducted sensitivity analyses using models with arcsine-transformed 26 iHOT-33 and HAGOS scores. The method for transforming the dependent variables is described in Appendix B.…”

Section: Discussionmentioning

confidence: 99%

Are cam morphology size and location associated with self‐reported burden in football players with FAI syndrome?

Scholes

Kemp

Mentiplay

et al. 2022

Scandinavian Med Sci Sports

View full text Add to dashboard Cite

Cam morphology size and location might affect the severity of reported burden in people with femoroacetabular impingement (FAI) syndrome. We investigated the relationship between cam morphology size (i.e., alpha angle) and selfreported hip/groin burden (i.e., scores for the International Hip Outcome Tool-33 (iHOT-33) and Copenhagen Hip and Groin Outcome Score (HAGOS)), examined separately for the anteroposterior pelvis (AP) and Dunn 45° radiographs in football players with FAI syndrome. In total, 118 (12 women) subelite football (soccer or Australian football) players with FAI syndrome with cam morphology (alpha angle ≥60°) participated. One blinded assessor quantified superior and anterosuperior cam morphology size by measuring alpha angles for the AP and Dunn 45° radiographs, respectively. Linear regression models investigated relationships between alpha angle (continuous independent variable, separately measured for the AP and Dunn 45° radiographs) and iHOT-33 and HAGOS scores (dependent variables). Larger anterosuperior cam morphology (seen on the Dunn 45° radiograph) was associated with lower (i.e., worse) scores for the iHOT-Total, iHOT-Symptoms, iHOT-Job, and iHOT-Social subscales (unadjusted estimate range −0.553 to −0.319 [95% confidence interval −0.900 to −0.037], p = 0.002 to 0.027), but not the iHOT-Sport (p = 0.459) nor any HAGOS scores (p = 0.110 to 0.802). Superior cam morphology size (measured using the AP radiograph) was not associated with any iHOT-33 or HAGOS scores (p = 0.085 to 0.975). Larger anterosuperior cam morphology may be more relevant to pain and symptoms in football players with FAI syndrome than superior cam morphology, warranting investigation of its effects on reported burden and hip disease over time.

show abstract

Section: Discussionmentioning

confidence: 99%

Are cam morphology size and location associated with self‐reported burden in football players with FAI syndrome?

Scholes

Kemp

Mentiplay

et al. 2022

Scandinavian Med Sci Sports

View full text Add to dashboard Cite

show abstract

“…26 As the iHOT-33 and HAGOS are self-reported measures that quantify patients' perceptions of their hip/groin burden, scores may also be influenced by nonphysical (eg, psychological, social, contextual) factors. 10,65 Nonphysical factors might explain more of the variance in iHOT-Total a 63 [23] 62 [23] 66 [20] iHOT-Symptoms a 70 [23] 69 [24] 72 [20] iHOT-Sport a 45 [27] 46 [31] 44 [18] iHOT-Job b 72 [34] 72 [38] 71 [28] iHOT-Social a 63 [32] 62 [32] 65 [31] HAGOS-Symptoms a 61 [14] 61 [14] 59 [23] HAGOS-Pain a 75 [18] 75 [18] 75 [20] HAGOS-ADL a 80 [20] 80 [20] 80 [25] HAGOS-Sport a 66 [25] 63 [25] 66 [26] HAGOS-PA c 63 [38] 63 [38] 63 [28] HAGOS-QOL a 60 [20] 60 [25] 60 [16] ADL, activities of daily living; FAI, femoroacetabular impingement; HAGOS, Copenhagen Hip and Groin Outcome Score; iHOT-33, International Hip Outcome Tool-33; IQR, interquartile range; PA, participation in physical activity; QOL, quality of life. reported burden than imaging findings alone and warrant further investigation.…”

Section: Discussionmentioning

confidence: 99%

“…Arcsin transformation of the dependent variables (PROM scores) can be used to stabilize variance and minimize bias in models. 31 Sensitivity analyses using models with arcsin-transformed PROM scores are described in Appendix B (available online).…”

Section: Methodsmentioning

confidence: 99%

Does Femoroacetabular Impingement Syndrome Affect Self-Reported Burden in Football Players With Hip and Groin Pain?

Scholes¹,

Kemp²,

Mentiplay³

et al. 2022

Sports Health: A Multidisciplinary Approach

View full text Add to dashboard Cite

Background: It is unknown if football players with femoroacetabular impingement (FAI) syndrome report worse burden than those with other causes of hip/groin pain, and to what extent this is mediated by cartilage defects and labral tears. Hypothesis: Football players with FAI syndrome would report worse burden than other symptomatic players, with the effect partially mediated by cartilage defects and/or labral tears. Study Design: Cross-sectional study. Level of Evidence: Level 4. Methods: Football (soccer and Australian football) players (n = 165; 35 women) with hip/groin pain (≥6 months and positive flexion–adduction–internal rotation test) were recruited. Participants completed 2 patient-reported outcome measures (PROMs; the International Hip Outcome Tool–33 [iHOT-33] and Copenhagen Hip and Groin Outcome Score [HAGOS]) and underwent hip radiographs and magnetic resonance imaging (MRI). FAI syndrome was determined to be present when cam and/or pincer morphology were present. Cartilage defects and labral tears were graded as present or absent using MRI. Linear regression models investigated relationships between FAI syndrome (dichotomous independent variable) and PROM scores (dependent variables). Mediation analyses investigated the effect of cartilage defects and labral tears on these relationships. Results: FAI syndrome was not related to PROM scores (unadjusted b values ranged from −4.693 ( P = 0.23) to 0.337 ( P = 0.93)) and cartilage defects and/or labral tears did not mediate its effect ( P = 0.22-0.97). Conclusion: Football players with FAI syndrome did not report worse burden than those with other causes of hip/groin pain. Cartilage defects and/or labral tears did not explain the effect of FAI syndrome on reported burden. Clinical Relevance: FAI syndrome, cartilage defects, and labral tears were prevalent but unrelated to reported burden in symptomatic football players.

show abstract

“…A total of four models using six different estimation methods were explored. The statistical models chosen in this study were based on previous literature that has tested the alternative methods and took into account the possible non-normal distribution of the data [8,12,15,17,19]. The best model was chosen based on the lowest RMSE and MAE.…”

Section: Discussionmentioning

confidence: 99%

“…A common characteristic of health utilities is the non-normal distribution and the upper bound value of 1. To address this, researchers have shown a potential benefit in transforming health utility responses prior to fitting it with a linear regression model [12]. Alternatively, other models such as the generalised linear model, Tobit and Censored Least Absolute Deviation (CLAD) have been explored [13,14].…”

Section: Introductionmentioning

confidence: 99%

Mapping PedsQL™ Generic Core Scales to EQ-5D-3L utility scores in transfusion-dependent thalassemia patients

et al. 2021

View full text Add to dashboard Cite

Purpose To develop a mapping algorithm for generating EQ-5D-3L utility scores from the PedsQL Generic Core Scales (PedsQL GCS) in patients with transfusion-dependent thalassemia (TDT). Methods The algorithm was developed using data from 345 TDT patients. Spearman's rank correlation was used to evaluate the conceptual overlap between the instruments. Model specifications were chosen using a stepwise regression. Both direct and response mapping methods were attempted. Six mapping estimation methods ordinary least squares (OLS), a log-transformed response using OLS, generalized linear model (GLM), two-part model (TPM), Tobit and multinomial logistic regression (MLOGIT) were tested to determine the root mean squared error (RMSE) and mean absolute error (MAE). Other criterion used were accuracy of the predicted utility score, proportions of absolute differences that was less than 0.03 and intraclass correlation coefficient. An in-sample, leave-one-out cross validation was conducted to test the generalizability of each model. Results The best performing model was specified with three out of the four PedsQL GCS scales-the physical, emotional and social functioning score. The best performing estimation method for direct mapping was a GLM with a RMSE of 0.1273 and MAE of 0.1016, while the best estimation method for response mapping was the MLOGIT with a RMSE of 0.1597 and MAE of 0.0826. Conclusion The mapping algorithm developed using the GLM would facilitate the calculation of utility scores to inform economic evaluations for TDT patients when EQ-5D data is not available. However, caution should be exercised when using this algorithm in patients who have poor quality of life.

show abstract

Exploring the Benefits of Transformations in Health Utility Mapping

Cited by 3 publications

References 27 publications

Are cam morphology size and location associated with self‐reported burden in football players with FAI syndrome?

Are cam morphology size and location associated with self‐reported burden in football players with FAI syndrome?

Does Femoroacetabular Impingement Syndrome Affect Self-Reported Burden in Football Players With Hip and Groin Pain?

Mapping PedsQL™ Generic Core Scales to EQ-5D-3L utility scores in transfusion-dependent thalassemia patients

Contact Info

Product

Resources

About