Rasch Measurement and Patient Reported Value Claims: A Primer for Hemophilia

Abstract

This commentary proposes that Rasch Measurement Theory (RMT) is an innovative method for assessments of patient-centric therapy response in hemophilia A and B, as they are in other disease states or target patient populations. RMT is a necessary and sufficient approach to moving from ordinal observations to interval measurement, which has arithmetic properties. This applies across the board in hemophilia and other disease states for clinical value claims, patient centric or subjective value claims as well as those for anticipated drug utilization and other medical care resources. The purpose of this commentary is to point out limitations regarding current methods for making claims regarding hemophilia response and to propose a new start in hemophilia studies to identify core claims that meet required measurement standards. This applies to both the development of new patient reported outcome instruments as well as the evaluation of existing instruments, with a focus on polytomous instruments and their sub-domains, to evaluate their possible application as measures that approximate RMT requirements.

Reliability and construct validity of the ACTIVLIM‐Hemo and Haemophilia Activities List (HAL) questionnaires in individuals with haemophilia

“…Given its intrinsic psychometric properties, 17 ACTIVLIM‐Hemo delineates a more homogeneous construct than HAL, which also facilitates the clinical interpretation of the patient's functional status. Altogether, ACTIVLIM‐Hemo is a measure of activity limitations in PwH that fills the lack of instruments that meet the minimum standard for measurement as claimed by Langley 6 . With its unique psychometric properties, ACTIVLIM‐Hemo is a very promising patient centric instrument to assess activity limitations in haemophilia, especially in PwH with access to regular haemophilia treatments.…”

“…Conversely, the HAL reports ordinal scores that are known for their lack of linearity and impedes the measurement of change. [4][5][6][7][8][9] Also, since the measurement error is only known on average for the HAL, the clinical interpretation of change scores further complicates the assessment of true significant change for a single patient. This stems for a restricted use of the HAL as a cohort descriptive assessment tool while the ACTIVLIM-Hemo, being built with a range-specific known discrimination, offers a larger potential to assess responsiveness to treatment across its measurement range.…”

“…This probabilistic formulation can then be used to determine 1 the level of activity limitations of each PwH and 2 the level of difficulty (or the weight) of each item. According to Langley, 6 the Rasch model is the only model providing the necessary and sufficient transformations, if possible, between ordinal observations and interval measures of patient‐reported outcome or cost effectiveness claims.…”

“…According to Langley,6 the Rasch model is the only model providing the necessary and sufficient transformations, if possible, between ordinal observations and interval measures of patient-reported outcome or cost effectiveness claims. At present time, activity limitations in PwH are typically assessed with the Functional Independence Score in Haemophilia (FISH) and…”

“…Conversely, the ACTIVLIM-Hemo was built from activities selected to delineate a unidimensional construct, where each item contributes complementary information of the same nature to determine the level of activity of PwH. In clinical practice, since the ACTIVLIM-Hemo is a unidimensional scale, a patient change affects the odds for the patient to succeed any item of the questionnaire by the exact same amount (actually the odds of passing any given level of difficulty along the linear scale increases by the same amount all along the scale) [6][7][8][9]. The more homogeneous nature of ACTIVLIM-Hemo compared to the HAL is also supported by the results of the multiple linear regression.…”